Skip to main content

Fast Company

AI is outpacing guardrails. Businesses must build their own

1 month ago

Dario Amodei, the CEO of Anthropic, recently joined the chorus of Silicon Valley Cassandras calling on AI companies to deliberately slow the pace of product development. In a widely discussed letter, he predicted a nightmare scenario in which a swarm of AI agents takes over the internet and causes hundreds of billions of dollars in damage—something he thinks could happen within the next six to 12 months.

But we don’t have to wait a year to see the damage artificial intelligence can cause if ungoverned. Most near-term harm will come not from rogue agents but from ordinary systems, bought from ordinary vendors, making ordinary decisions about people. For example, in the banking industry, AI is already helping decide who gets a loan and at what price, whose claim is flagged, and whose account is closed. According to Deloitte, twice as many bank employees use AI weekly as did a year ago and the number of publicly reported AI incidents and hazards in the sector was eight times higher in 2025 than in 2022.

That might not feel like an existential problem, but it has real consequences for people’s finances and well-being—and the same problem exists across health care, transportation, and other critical sectors.

The labs may be able to pace what they build, but not what the rest of us deploy. Nor are regulators moving fast enough to protect your business, your customers, or your shareholders’ interests. Every business must decide for itself what it will and will not let a machine decide.

A framework to fill the regulatory vacuum

The potential consequences of ever-increasing AI uptake plus lagging governance are clear. An agentic system connected to internal infrastructure can open the door to cyberattack. An employee untrained in AI can inadvertently perpetuate hallucinations or misinformation. An underwriting algorithm based on biased historical data can quickly deny qualified borrowers.

Real-world impacts like these require more than a page of AI use guidelines that gets posted on an intranet and never referenced again. Over the last year, I’ve worked with 14 banks across the Global Alliance for Banking on Values to build and test a comprehensive framework to guide AI governance in this regulatory vacuum. While the need is urgent in banking, this framework applies equally to other critical sectors, including health care, insurance, retail, and telecommunications, where AI-enabled decision making intersects with human lives.

Business leaders should start by acknowledging that AI needs guidance in a way other technologies do not. AI decides rather than implements. It reasons, but that reasoning is a black box. For that reason, AI governance can’t just be another risk register where your board reviews a list of things that can go wrong and how likely that is to happen. AI governance requires business leaders to contend with the good a business has failed to do and the customers who were never reached because the model failed to see them. 

In practical terms, there are three immediate steps that business leaders can take across sectors to start building that moral architecture, all revolving around amping up human oversight at every step:

1. Name an owner for every AI system

Start with an inventory of where AI is doing the work. You cannot govern what you have not listed, and institutions often find more AI embedded in their work than expected. Plenty of AI enters a company without ever passing through model review, arriving instead through vendor software or tools employees adopt on their own. So the first step is a complete inventory of every AI system operating in the institution, no matter how it got there.

Each system then needs a named owner who has real authority. That person must have the standing to enact changes, responsibility for periodically revalidating the system, and, if necessary, the ability to take the system offline.

2. Keep human judgment in the loop

Hospitals are increasingly using AI to facilitate diagnoses, but a qualified medical expert is always the one to make the final clinical decision and explain it to the patient. The same must go for banking or any other business. Employees using AI systems need training to evaluate what those systems produce and feel empowered to override outputs. They must also be able to explain the final decision to both customers and regulators in plain language, without an engineer translating.

Track how often employees override AI systems. If a system has run for a year and no one has ever disagreed with it, human judgment is lacking.

3. Include vendors in your governance strategy

Most institutions buy AI rather than build it, so governance must travel upstream into every contract. Fully 91% of senior executives in a recent survey said that they don’t fully understand their organization’s dependencies across AI vendors, models, and infrastructure. The concentration on a few foundation models is such that Moody’s recently raised the alarm that if just one of those providers has an outage or raises prices, the effects could ripple across the entire banking sector—a risk that applies to other sectors as well.

Make sure to embed vendor accountability into your AI governance frameworks. That means including things like audit rights, bias-testing support, incident notification, and the right to terminate in all vendor contracts.

It’s not all doom and gloom

Pessimistic predictions about AI make headlines, but the optimistic vision is the strongest argument for taking action. AI has the potential to reduce customer friction, make space for employees to take on more meaningful work, and expand economic opportunity. For example, data across more than 180 banks and credit unions using Zest AI’s machine learning model—which uses alternatives to traditional credit scores to assess risk—shows 49%, 41%, and 40% increases in loan approval rates for Latino, Black, and women borrowers, respectively. Far from being a tax on innovation, proper AI governance can help broaden market share and reputational value.

None of that happens automatically. When regulators do eventually catch up, they’ll either be codifying what responsible companies have already built or cleaning up after the ones that waited to be told. Businesses get to decide right now which of those they’ll be, and every one of us has a stake in the answer.

Half of the world’s population has been exposed to heat waves and wildfire smoke at the same time

1 month ago

Climate change is a global phenomenon, meaning that everyone, in some way, will feel its devastating impacts. 

And increasingly, those impacts are happening simultaneously: Over the past two decades, more than half of the world’s population has experienced simultaneous heat waves and wildfires, occurring on the same day, a study found.

Researchers at Monash University in Melbourne, Australia, published their findings this week in the journal Science Advances.

The study looked at 20 years of data, from 2000 to 2019. To see who experienced a wildfire, researchers looked at who was exposed to “fire-sourced air pollution”; their data took into account hourly temperature, fire emissions, air quality, and population information.

Over those 20 years, researchers found, the number of people around the world exposed to both wildfires and heat waves, on the same days, increased by 287 million—from 761 million people per year in the first half of that time period to more than 1 billion people per year in the second.

That means 55.9% of the global population has experienced at least one day where a heat wave and wildfire pollution occurred in tandem.

And the number of such days is increasing. From 2000 to 2009, the average person experienced one such day per year. Between 2010 and 2019, that rose to 1.7 days of co-exposure per person.

Damaging on multiple levels

The research doesn’t only illustrate how climate change is making extreme weather more frequent. It also shows how the health impacts of these events can compound as extreme weather overlaps.

Both heat waves and wildfires can cause serious harm to our health. Along with the risk of heatstroke—which can damage your brain, heart, and kidneys, and even be fatal—extreme heat and humidity can make it harder for some people to breathe, especially those with asthma or other lung conditions. 

Wildfires come with intense air pollution, full of tiny particles called PM2.5 (particulate matter 2.5 micrometers or less in diameter) that can get into our lungs and bloodstream, raising the risk of heart attack, stroke, and reduced lung function.

When heat waves combine with air pollution, the health impacts exacerbate. On days when an area sees a “double whammy” of high heat and high air pollution, the risk of death is three times greater “than the effect of either high heat or high air pollution alone,” health researchers have previously written. 

The Monash University researchers emphasize this in their study, too. “Compared with exposure to heat wave or landscape fire-sourced air pollution alone, co-exposure to both hazards may exert greater health impacts than the sum of their individual effects,” they write. 

In other words, the health risks—including respiratory and cardiovascular impacts—become amplified when wildfire air pollution and extreme heat happen together.

Extreme heat and wildfires also impact each event occurring. Conditions like high temperatures and dry soil promote wildfires, and wildfires then exacerbate heat waves because they release heat and pollution into the air.

“These processes create a positive feedback loop in which heat waves and fires amplify one another,” the researchers wrote.

Which regions are being hit hardest?

Some regions of the world are experiencing these simultaneous effects more than others. Zambia, for example, sees more than 28 such days per person every year. The Democratic Republic of the Congo comes in second, with just over 24 days per person per year.

Countries with lower life expectancy, education, and income tended to see the highest amount of wildfire and heat co-exposure, an example of the way that climate change is hitting the world’s poorest countries the hardest. 

Still, almost every area of the world saw an increase in this trend.

Central America, South America, Central and Eastern Africa, Southeast Asia, and parts of North America, Australia, and Russia all experienced “notable increases” in the annual average number of days that wildfire pollution and extreme heat occurred together.

Wholesale inflation reaches levels last seen in 2022 amid soaring energy costs

1 month ago

Wholesale inflation rose more than expected last month on a big jump in fuel costs as rising prices become a bigger economic challenge for millions of Americans, and a growing political problem for Republicans in the upcoming midterm elections.

The Labor Department’s producer price index—which captures inflation before it reaches consumers—soared 6.3% in September from a year ago, up from 5.6% in August and the biggest increase since December 2022, the government said Thursday.

On a monthly basis, wholesale prices rose 1.3% from August to September, marking the largest increase since March 2022, after a 0.5% gain the previous month.

The data indicated that rising fuel prices, particularly for the diesel fuel used by long-haul trucks and freight trains, is pushing up the cost of shipping and warehousing goods.

That is starting to lift the cost of food, one place where prices appeared to have cooled this year. Wholesale food prices jumped 1.2% just from August to September, though they are still unchanged from a year ago.

Transportation and warehousing prices climbed 3.4% last month and are 16% higher than a year ago, the report showed.

“The jolt to transportation prices affects every good in the economy, since inputs to manufacturing move to factories and goods at all stages of production move from producer to wholesaler to retailer by truck, train, or rail,” Carl Weinberg, chief economist at High Frequency Economics, wrote in a client note. “This is how increases in energy prices ‘bleed through’ to prices of core goods.”

Farmers also rely on diesel to power most of their equipment, which could also push up food prices. Diesel prices averaged $6.32 a gallon nationwide Thursday, according to AAA, up from just $3.66 a year ago.

Thursday’s report amplifies the consumer price index released Wednesday, which also accelerated in September as gas prices jumped. But a broad range of goods and services became more expensive.

Both inflation reports raise pressure on the inflation-fighters at the Federal Reserve to hike their key interest rate. Inflation has remained above the central bank’s 2% target for five and a half years.

Excluding the volatile food and energy categories, core prices rose 0.7% from August to September, up sharply from 0.2% in the previous month. Compared with a year ago, core prices rose 5%, up from 4.6% in August.

More expensive energy drove most of the increase as wholesale gas prices soared more than 20% from August to September. But the report also showed other costs rising, including those for electricity, vegetables, shipping, and airfares.

Ongoing inflation has eroded Americans’ incomes and soured most consumers on the economy. The Labor Department said Wednesday that for the first time since May, average weekly paychecks when adjusted for inflation shrank compared with the same month last year.

Higher prices have taken a toll on Trump’s standing with U.S. voters. A poll released this month by The Associated Press-NORC Center for Public Affairs Research found that just 17% of U.S. adults approve of Trump’s handling of the cost of living. And only 26% approve of his handling of the economy overall, a new low.

The Federal Reserve, meanwhile, will hold a policy meeting in two weeks and may keep its key short-term interest rate unchanged. After two hot inflation reports this week, however, some economists now expect that it could hike rates. Officials have penciled in at least one more rate hike this year.

Fed Chairman Kevin Warsh speaks at an international economic conference late Thursday in Bangkok and his remarks will be closely watched for any hints of whether he supports a rate hike this month.

The Fed uses higher rates to slow borrowing and spending in a bid to lower inflation.

That has not proved effective as hoped this year.

After a rate hike in September, the U.S reported Thursday that retail sales rose faster than expected during the same month. A chunk of that spending, however, was at gas stations, where Americans are paying more to fill up their cars.

—By Christopher Rugaber, AP economics writer

A timeline of New York City trying to ban horse carriages in Central Park

1 month ago

The iconic but controversial horse-drawn carriage industry in New York City might finally come to an end as Mayor Zohran Mamdani and New York City Council Speaker Julie Menin reached a deal Tuesday to outlaw them.

If enacted, the bill would ban rides for hire starting June 1, 2028. 

Other NYC mayors have contributed to this effort, but none facilitated a full ban. Mamdani could be the first to achieve the full support of the City Council to effectively ban horse-drawn carriages.  

How prior mayors dealt with carriage controversies

When a horse collapsed from exhaustion in 2011, then-Mayor Michael Bloomberg dismissed a call to ban the industry, citing the importance of maintaining a livelihood of New York City’s carriage drivers.

In 2013, former Mayor Bill de Blasio was the first mayor to announce plans for a full ban. Without support from the City Council, however, the effort proved fruitless.

Just months before the end of his mayorship, de Blasio’s successor, Eric Adams, attempted to end horse-drawn carriage operations, which also lost due to a lack of support from the City Council. 

Why is Mamdani acting now?

An 18-year-old tourist from India, Romanch Mahajan, died after falling from a runaway carriage in July, which prompted “Romanch’s Law.”

The legislation will phase out the carriages while also providing a plan for protection of the workers whose livelihoods rely on the popular tourist activity. This time around, council members are expected to support the ban, after Menin publicly endorsed the law in July.

“The horse carriage industry is coming to an end in New York City. Our administration will make sure workers and owners are protected and the horses are safe,” Mamdani said in a statement on X on Tuesday. 

What’s at stake for the industry?

The industry salary is estimated at $65,000 to $70,000 per year, with approximately 68 active licenses and just under 160 horses licensed at the end of last year, according to the City Council’s Committee on Health.

As of May 3, 2025, the maximum fare for a carriage ride is $72.22 for the first 20 minutes, plus $28.89 for each additional 10 minutes. 

The 163-year-old industry brings tourism dollars into the city, although current estimates for just how much are hard to come by. (We’ve asked the NYC Horse & Carriage Alliance for more details.)

But it’s also yielded unsafe conditions for the horses and passengers. There have been 40 horse deaths between 1982 and 2026, the Manhattan borough president’s office found.

Central Park South in New York City circa 1965. [Photo: John Garetti/Getty Images]

As the city moves toward a ban, the mayor is taking workers into consideration in the next stages.

CarriageOn, a coalition of carriage operators, cites itself as a diverse workforce, with immigrants at the forefront of its operation. There are questions about the future of these carriage operators, which the Mamdani administration aims to address.

The mayor’s office has been working on a detailed plan for the transition of the workforce and industry, a spokesperson on Mamdani’s team told Fast Company. 

Why does NYC have the horses in the first place?

Horse-drawn carriage rides for tourists date back to when Central Park opened in the 1800s, when carriages were already a mode of everyday transportation. Even then, people cited concerns about the horse-drawn vehicles. 

One of the most outspoken organizations is the American Society for the Prevention of Cruelty to Animals (ASPCA), which was founded in New York City in 1866 amid a growing concern about the well-being of horses in the city.

Its original logo brandished an image of a carriage operator abusing a horse, while an angel hovered over the animal. Cruelty toward horses was a driving factor for the organization’s foundation, but horses also remained a vital part of transportation in the city. 

What would a ban mean for the horses?

As the tradition phases out, the horses need somewhere to go.

“If Romanch’s Law is enacted, the ASPCA stands ready to support the City through this transition,” ASPCA president and CEO Matt Bershadker told Fast Company. The organization is ready to provide “grant funding or other resources if necessary to help ensure humane placement and a safe future for every horse in need of a home,” he added.

Speaker Menin announced on Wednesday that the full council will vote on October 22 on the legislation.

“We have been working closely with the City Council, union partners, carriage drivers, animal welfare advocates, and community leaders to deliver a just transition that protects workers while ending horse-drawn carriages in Central Park once and for all,” Sneha Choudhary, deputy press secretary for the mayor’s office, told Fast Company.

We can’t keep guessing whether AI helps students learn

1 month ago

On September 2, as students returned to class, New York City implemented a one-year pause on generative AI use for nearly 600,000 students across pre-K, elementary, and middle school. Los Angeles soon followed, blocking it for every student on a district-issued device. Together, the country’s two largest school systems came to the same conclusion: Better to wait a year than to keep putting untested tools in front of kids.

Neither district is calling the decision permanent. Both are responding to parents who worry that untested AI tools are reaching students before anyone knows whether they improve learning. New York City Schools Chancellor Kamar Samuels said it best: “We’re standing firmly in our belief that innovation does not mean more technology, and over the next year, we will lead with evidence to make sure technology serves learning—not the other way around.” 

That leaves an obvious question: Where will that evidence come from?

ADOPTION IS OUTPACING EVIDENCE

Teachers, school leaders, parents, and students are making consequential decisions about AI without reliable information about what works, for whom, and under what conditions.

Sixty percent of teachers use AI for work, yet 82% report they’ve received no formal guidance on how to use it. Just half of middle and high schools have any AI policy, and only 6% of teachers in those schools say those policies are clear. This lack of clarity shows up in how people feel about AI, too. Thirty-one percent of Gen Z now say AI makes them angry, up from 22% a year earlier, and 80% believe relying on it will make learning harder.

Still, banning AI in school won’t take it out of students’ lives completely. A recent Common Sense Media survey found that 91% of teens ages 13 to 15 use AI for schoolwork or homework. And with 21% of adults already using AI at work, students need to consider the growing role of AI in their future careers.

It makes sense that school district leaders, educators, and families want to know when AI adds to student learning and when it takes away from it. But even with the best intentions, it’s hard to write good policy, train teachers, or make smart purchasing decisions when the evidence underneath is this thin.

WHAT WASHINGTON SHOULD DO

Congress and the federal administration have a chance to put the existing infrastructure to work and provide the evidence we need to safely and confidently advance student learning. The Alliance for Learning Innovation worked with AI researchers, practitioners, and our coalition of more than 150 members to develop federal recommendations on this exact topic. Four main takeaways:

  1. Pass the NEED Act. This past September, Senators Michael Bennet (D-CO), Thom Tillis (R-NC), and Chris Van Hollen (D-MD) reintroduced the bipartisan NEED Act, which would create a national center at the Institute of Education Sciences (IES) to develop and test bold new approaches to teaching and learning.

The model comes from the Pentagon’s Defense Advanced Research Projects Agency (DARPA), which helped build the early internet and laid the groundwork for Siri. Like DARPA, the new center would be run by expert program managers who can quickly move money to promising and high-impact ideas, invest more when they show results, and pivot when they don’t.

  1. Study AI tools with robust data. One example of how to study the issue: New York City plans to run AI pilots in a small number of high school classrooms this year to learn what’s working for its own students. The federal government should help school districts nationwide do the same.

Where pilots aren’t feasible, a coordinated national network of testbeds, combining real classrooms with synthetic environments that simulate how students use AI, would let researchers test tools across diverse schools and communities. The result would be practical evidence about how AI holds up in the everyday reality of teaching, plus clearer signals for developers and investors about which tools deserve to scale. Federal investment in these testbeds would give schools a shared, safe way to judge what is and isn’t working in classrooms.

  1. Measure what students learn. Many current evaluations can’t tell the difference between a student’s output and AI. Federally funded AI studies should measure how students perform with a tool, then evaluate them again on the same topics later without it. The American Psychological Association made the same case, urging that any AI-powered learning tool be judged by what students can do after they stop using it.

This is exactly the kind of rigor the Department of Education’s own guidance and a growing number of outcomes-based contracting models (which tie vendor payment to results), are asking for. Currently, few evaluations consistently meet this standard.

  1. Give school districts a common evaluation tool. Districts are evaluating AI tools without a clear definition of what makes it safe, effective, and ethical. The National Institute of Standards and Technology (NIST) already published a voluntary framework for AI Risk Management that organizations use across industries. IES should partner with NIST to build an education-specific version, aligned with the EDSAFE AI Alliance’s SAFE Benchmarks, that gives states and districts shared standards for judging tools on learning impact, safety, privacy, and transparency before those tools reach classrooms.
THE CLOCK IS TICKING

New York City was clear that this moratorium lasts one year for a reason. The city said it will spend that time studying how AI affects students, and Los Angeles and other districts have set similar deadlines.

That study shouldn’t happen one district at a time, and it shouldn’t be limited to the contexts of large and urban school districts. If we invest in our federal research infrastructure now, we can build a trustworthy evidence base across the country so that all students walking into class in Fall 2027 have access to tools that we know work.

Sara Schapiro is executive director of the Alliance for Learning Innovation.

Dementia cases are expected to more than double by 2050. A new study reveals who’s most at risk

1 month ago

More and more of our loved ones could develop dementia, as the number of diagnoses is projected to more than double by 2050, according to new research.

In the absence of large-scale prevention efforts, the number of people who are diagnosed with dementia each year could total 24.7 million by 2050, up from 11.4 million in 2023, according to a study by Australian researchers published in JAMA Neurology. Women and people who live in low- and middle-income countries already have higher rates of dementia, a trend that’s likely to continue, according to the findings.

While the researchers didn’t pinpoint an exact reason why dementia affects populations differently, they say the prospect of more cases warrants urgency in investing in prevention efforts. The prospect of 24 million new cases of dementia each year would place “enormous pressure on individuals, families, health systems and economies around the world,” Henry Brodaty, co-director of the Centre for Healthy Brain Ageing at the University of New South Wales Sydney and a co-author of the study, said in a statement.

“We cannot simply wait for people to develop dementia and then respond,” Brodaty said. “We need to act across the spectrum—from reducing modifiable risk factors and identifying people at risk earlier, through to improving diagnosis, treatment and care.”

The researchers studied data from more than 38,000 people diagnosed with dementia spanning 18 countries or territories on six continents. They found that the incidence was somewhat higher among women—at 10.6 cases per 1,000 people versus 8.6 for men—but the difference was even starker depending on where people lived. The incidence of dementia was 13.6 cases per 1,000 people in low- and middle-income countries, compared with 8.3 in high-income countries.

Understanding the populations who are most at-risk for developing dementia will be key to developing prevention, diagnosis and care strategies, according to the study’s authors. 

“The higher incidence we observed in low- and middle-income countries is particularly concerning because these regions are experiencing rapid population aging while often having the fewest resources for prevention, diagnosis and care,” Perminder Sachdev, co-director of the Centre for Healthy Brain Ageing at the University of New South Wales in Sydney and a senior author of the study, said.

THE FINANCIAL TOLL OF DEMENTIA

Aging populations are understandably one of the main reasons why dementia cases are expected to increase, as this condition that affects memory, thinking and the ability to perform daily activities is more common after the age of 65. As of 2021, an estimated 57 million people worldwide were living with dementia, according to figures from the World Health Organization, and it’s the seventh leading cause of death.

The World Health Organization likewise estimated that as of 2019, dementia cost economies around the world about $1.3 trillion, so a surge in new cases also carries a financial toll. 

That’s why the new research should sound an alarm, as Dr. Liron Sinvani, director of research and innovation at the Northwell Institute for Healthy Aging and a professor at the Feinstein Institutes for Medical Research in New York, told U.S. News & World Report. “It’s an accelerating epidemic that demands immediate attention.” 

And that’s especially important as the countries that will face an “enormous wave” of dementia cases also have very limited infrastructure to respond, as Sinvani told the outlet. 

“Health systems must prepare for this massive increase in diagnostic and care demand, and this is especially true for these low- and middle-income countries where the growth will be fastest,” she said. “We know that dementia care requires a lot of caregiver support and is very costly. We will need a better infrastructure.”

Social Security 2027: VA disability payments to increase alongside COLA. Here’s how much the adjustment amounts to

1 month ago

Some good news for military veterans who collect disability and retirement benefits each month: Effective December 1, military retirement checks and Department of Veterans Affairs disability benefits will increase 3.6%. (However, veterans will most likely not see the increase until December 31 or early January 2027, according to Stars and Stripes.)

The increase is tied to the Social Security Administration’s announcement on Wednesday of the largest annual cost-of-living adjustment (COLA) in years, for next year, 2027. That number was based on the Bureau of Labor Statistics’ (BLS) Consumer Price Index data from September inflation data, showing inflation at 3.7%, up from 3.4% in August.

Like other Americans who receive Social Security benefits, those receiving veterans’ disability and retirement checks will also get a 3.6% bump, which amounts to an extra $36 for every $1,000 in their pension, per month.

According to Robert Laura, a financial planner and founder of the Retirement Coaches Association, “Many military retirees are far younger than the typical Social Security recipient, so they’re still paying a mortgage, raising kids, supporting a spouse whose career took a hit from years of moves.” So, he says, “a 3.6% [increase] is welcomed, but it doesn’t come close to the rising costs many are carrying in their civilian rucksack.”

However, there is a silver lining.

“Veterans should expect to see a significantly greater impact from the COLA increase than non-veterans,” Greg Welborn, principal and financial advisor at First Financial Consulting, tells Fast Company. “[Their] benefits are tax-free so they pay no income tax on the increase; and they [also] won’t see any increase in Medicare premiums to offset the COLA.”

More than 8 million veterans in the U.S. receive disability and retirement benefits, according to Stars and Stripes.

Unlike Americans who receive military pensions and Social Security payments—which are automatically adjusted each year based on the COLA—VA disability benefits are subject to the Veterans’ Compensation Cost-of-Living Adjustment Act, which must be passed by Congress and signed into law by President Donald Trump.

Government has perfected awarding grants

1 month ago

The federal government sent $1.26 trillion to state and local governments in FY2022, primarily through grants. States distributed $662 billion to local governments, while nonprofits received $294 billion in government grants that year.

Grantmaking is essential to turning public policy into local action. Today, state agencies use grants to equip organizations to do everything from improving rural health and preventing homelessness to expanding broadband access and spurring economic development.

Most agencies already have a system for the transactional side of moving funds. There’s software for tracking the applications, awards, drawdowns, and procurement. The financial machinery of a grant is mature. But it wasn’t built to answer the question that matters most: Is the money working?

Beyond transactions and compliance, that’s a question about grantee needs and performance.

TODAY’S SYSTEMS SUPPORT THE GRANT

The answers to whether grants are working are scattered across spreadsheets, inaccessible shared drives, and a handful of annual reports.

But grantees are people, often working at community organizations, faith-based organizations, and other nonprofits, ranging from three employees to thousands. Today’s grant management systems don’t connect with those doing the work.

Does the executive director who inherits a grant midstream have what she needs? Can a frontline public servant get an answer in two days? When federal guidance shifts, is there any mechanism to broadcast the change to everyone who needs to know? Does the grantee have the organizational capacity to do this again, or is it being held together by one person who could leave?

We call the work of supporting grantees after an award “post-award delivery.” It exists to answer the question: Did it work?

GRANTEE SUPPORT IS AD HOC

A program officer’s actual job involves collecting the required documents, chasing outcome data on a cycle, and keeping some record of what was said to whom. The tools at their disposal weren’t built for the work. Contacts live in spreadsheets. Guidance goes out as an email attachment. And there’s no communication trail, just an inbox and whatever the last person remembers telling someone.

The cost isn’t evenly distributed. The best-resourced grantee asks the most questions, because someone there has “call the state” in their job description. The three-person nonprofit asks none because it doesn’t have the bandwidth to know what it doesn’t know, and it suspects admitting confusion won’t help next year’s application.

Tighter oversight doesn’t help here. Recipients need support. A grantee who can find an answer without waiting, who knows the agency is trying to make the job easier, is a grantee who trusts the process enough to ask for help before something breaks.

A COMPLEX CHAIN OF GRANTORS AND GRANTEES

State agencies are simultaneously grantees and grantors, often from the same desk in the same week. They receive federal funds with requirements attached, then distribute them to their partners with a layer of state requirements added.

Every fracture repeats at every tier. The state waiting on federal clarification has subrecipients waiting on it. Questions travel up the chain slowly and answers come back slower, and the organization at the bottom absorbs the accumulated latency of the entire structure.

BETTER METRICS TO UNDERSTAND GRANT EFFICACY

Today’s metrics measure the money’s movement. They’re process-oriented. We believe there are three metrics worth adding to understand grantee performance and improve program effectiveness.

  1. Agency throughput. How much time passes between appropriation and dollars doing work in the field, measured at the agency level across years rather than grant by grant?
  2. Capacity. Can this recipient run the program next cycle with less support than it needed this cycle?
  3. Cohort variance. The average grantee doesn’t exist. A cohort where 15 organizations are thriving and 20 are underwater reports the same average as one where everyone is merely muddling through. Aggregate numbers can’t tell you about the specific challenges.

With the right infrastructure, struggling grantee organizations can be clustered by size, region, or prior grant experience. Patterns emerge. Operations improve.

THE QUESTIONS ARE THE DATA

Recipients generate a continuous stream of intelligence that grantors currently discard.

Question volume and sophistication reveal a lot. If everyone is still asking foundational questions six months in, the program design has a problem. A cluster of questions from one region is an early warning. A recurring question is a documentation gap.

But today this data is functionally inaccessible. It’s scattered across inboxes, call notes, and untranscribed office hours. Structure it, and two things become possible:

  1. Routine questions get answered instantly against cited, approved guidance.
  2. Clustering surfaces emerging implementation problems while there’s still time to act.

The same infrastructure that lets grantors capture data also gives grantees a well-run program: a living repository of guidance, peers they can find, and one place to ask questions and see what others have already asked. Recipients who can see each other reduce the load on the agency.

It’s a shift from one-to-one to many-to-many. One answer, given once, available to every current and future recipient. Knowledge survives turnover, lasting beyond the period of performance.

BETTER GOVERNMENT FOR ALL

State agencies don’t want to be compliance enforcers. They want grantees to succeed and already invest enormous effort in helping them do so.

What’s missing is infrastructure that lets their guidance, expertise, and support reach every recipient when it matters. When agencies can identify challenges early, share answers once, and build capacity across a cohort, grantees perform better. Grants, in turn, deliver more for the communities they were designed to serve. Concretely, that means more veterans connected to housing and benefits, more communities connected to broadband, and more rural patients reached.

Getting post-award delivery right is the critical last mile between public funding and visible impact.

Madeleine Smith is cofounder and CEO of Civic Roundtable.

Americans are warming to a healthcare idea that once seemed politically impossible

1 month ago

Americans have hit a breaking point with health insurance.

Saddled with growing medical bills and totally reliant on stable jobs to stay healthy, the majority of Americans now say they would prefer a government plan that offers health insurance to everyone. 

Two-thirds of U.S. adults now support a national healthcare plan or “Medicare for All,” according to recent polling from health policy research group KFF. That’s a dramatic increase from 2020, when 53% of U.S. adults favored universal healthcare through a government plan, an option common among other wealthy nations. 

In the poll, 32% of respondents said they would “strongly favor” a single government healthcare plan that provides insurance for all Americans. An additional 34% of those polled said they would “somewhat favor” the overhaul to the U.S. healthcare system, with just 15% saying they “somewhat oppose” the idea and 19% saying they “strongly oppose” it. In the KFF poll, 84% of Democrats and 72% of independents were in favor of a national health insurance plan, compared with 34% of Republicans.

Another survey published this week by Axios-Ipsos confirmed the swell of support for reinventing the current U.S. healthcare system. In that poll, 43% of respondents said they would support a Medicare for All proposal that would replace the current private health insurance system. In that poll, 24% of respondents said they would oppose such a plan, while 32% reported being “unsure.” More Democrats (62%) and independents (47%) favored a government-run healthcare system, while a slight majority of Republicans opposed the idea (52%).

Crushing healthcare costs

Soaring support for a government-run health insurance system isn’t a mystery. Americans spend more on healthcare than people in other wealthy countries, but the U.S. government spends more, too—even without a broad national plan that insures everyone. Somehow all of that spending still adds up to worse health outcomes. In spite of spending twice what many of its peer countries spend per capita on healthcare, Americans live shorter, less healthy lives.

Americans are paying more and getting less when it comes to healthcare—and that spending is stretching them thin. In the richest country in the world by many measures, people are skipping doses of their meds and even buying less food to afford their rising healthcare costs. 

In the new Axios-Ipsos poll, 69% of Americans said they were worried about the rising cost of healthcare, with the same percentage reporting that a 10% increase in costs would hurt their family’s financial situation. A shocking 100 million Americans were already crushed under the weight of medical debt in 2022, and the situation is unlikely to improve given recent cuts to Affordable Care Act (ACA) subsidies. In 2024, Americans took out $74 billion in loans to pay their medical bills, with most borrowing more than $500.

U.S. voters fed up with the current state of health insurance will have a way to voice that displeasure soon. Progressive Democrats in some key November battleground races are running on a push for Medicare for All, but the party’s center and its leadership have yet to warm up to the idea in spite of soaring support among their constituents.

Your AI tools should be working together better

1 month ago

No one buys a dishwasher on Tuesday, then wakes up Wednesday thinking: “You know what I could really use? Another dishwasher.”

But that’s essentially how many e-commerce systems still operate. A shopper may spend weeks researching dishwashers—visiting a retailer’s website, using the mobile app, and eventually going to their local store to narrow down the options in-person—before making the big purchase.

Yet the next day, they return to the website, or open an email or text from the retailer, only to find themselves encouraged to buy…another dishwasher. It’s archaic, a little ridiculous, and often the result of a surprisingly simple problem: The systems that registered the shopper’s interest aren’t working from the same context as the one that registered the purchase.

What would be more useful is to recognize the customer’s needs have changed. Maybe now they need installation guidance, maintenance tips, detergent tabs, or something else complementary to what they just bought. But they almost certainly don’t need another dishwasher.

AI’S NEXT CYCLE

The first wave of GenAI showed what models can do with massive amounts of information, from answering natural-language questions (“Which dishwasher features matter most for a family of four?”) to summarizing complex materials to creating content.

But generating polished responses is different from understanding the big picture—including individuals’ actions, context, and the relationships between signals—and using that to determine what happens next.

To bridge that gap, there’s rising interest across industries in the concept of reasoning engines.

At a basic level, reasoning engines are AI systems that perform logical inference based on real-world inputs, not just questions or prompts people give them. And rather than considering one interaction or data point in isolation, they can reason across multiple signals and channels. This way, recommendations, including at the individual level, are most likely to land.

STRONG LIMBS, BUT NO BRAIN

Sprawling tech stacks have made the need for overarching reasoning more pronounced. This didn’t happen overnight. Organizations have spent years assembling and deploying specialized technologies, each designed to do one thing very well.

For example, in e-commerce, search interprets what a consumer is actively asking for (“a quiet dishwasher”), while recommendations predict and display products that might interest a given shopper. AI agents and chatbots interact with shoppers conversationally, and marketing systems manage communications across channels like email and SMS. In-store POS systems know what happened offline, including that the customer purchased a dishwasher. And so on.

AI is making these individual systems more capable. But greater intelligence at the point-solution level doesn’t automatically create a more intelligent experience overall.

Think of specialized systems as highly capable limbs. Strengthening each one doesn’t solve the coordination problem if there’s no central nervous system. In other words, if your left leg doesn’t know what your right is doing, turbo-charging them may just make you stumble faster.

That’s why a reasoning layer matters. It acts as the brain or connective intelligence, letting systems share context and coordinate decisions, so the whole body moves together.

REAL-WORLD REASONING

Across sectors, companies and their customers are starting to see the benefits of connected intelligence.

In banking, for instance, agents reason across risk signals, gauging what’s significant and what’s not, while reducing manual work across traditionally fragmented know-your-customer processes.

And in supply chains, companies are using central, connected intelligence to link planning, procurement, and logistics decisions, moving from “firefighting to orchestration.”

Retail offers another useful illustration, where innovative brands connect breadcrumbs across the shopper journey—from website and in-app behavior to loyalty program status—to inform what someone sees next. An online shopper checking out men’s Hawaiian shirts, for example, might ask the site’s AI assistant later that day for vacation outfits. Reasoning across the shopper’s browsing behavior and purchase history, the agent highlights men’s tropical gear in that person’s preferred size and colors.

DON’T START WITH THE TECH

These examples show some benefits early adopters are achieving. Others can, and should, begin laying the foundation.

Think about it. Companies built websites before e-commerce became ubiquitous. Similarly, they shouldn’t necessarily wait for reasoning technology to reach widespread adoption before building the architecture to support it.

Here are ways to get ready:

  • Break down context silos: Identify where important signals live and whether relevant systems can access and share them.
  • Whip underlying data into shape: A reasoning layer isn’t magically useful if the information it’s reasoning over is incomplete, inconsistent, or inaccessible.
  • Architect for orchestration: When adding AI agents or other applications, consider how they’ll connect with existing systems rather than evaluating them just on standalone capabilities.
  • Design around the user’s journey: Customers don’t experience channels and point solutions; they experience one brand. Technology architectures should reflect that.
TOMORROW’S MAINSTREAM CONCEPT

Reasoning tech isn’t new. It’s been part of the AI and data science lexicon for several years now. A 2022 Google Research paper introduced chain-of-thought prompting and voices like Yann LeCun, former chief AI scientist at Meta pursued a different architecture for how machines can reason and plan. Though largely still a technical concept, there’s precedent for how quickly that can change.

Flash back to 2017, for instance, when a Google Research paper introduced a new kind of neural network, the transformer. The concept was obscure then, but the paper has since been cited extensively. The architecture it described became the foundation nearly every modern LLM is built on—the “T” in ChatGPT.

Now, as AI makes individual solutions more capable, and the need for orchestration more important, reasoning engines may be on a similar trajectory.

And as we think back to that dishwasher, it’s clear why. That shopper doesn’t know or care which system powered which interaction. They just expect the brand to understand the context they’ve already provided. (Rinse and repeat!)

When enterprise tech behaves like that, we will have reached the next era of AI, where individual systems aren’t just smarter; they work together. Then, technology itself becomes less visible, and things simply work the way consumers expect they should.

Kevin Laymoun is chief customer officer and chief revenue officer at Constructor.

What happens when your employees’ AI agents start coming to work? (Spoiler: It’s bad)

1 month ago

Personal AI agents are having a moment. Meta recently launched a personal agent it named Muse and saw an estimated 2.8 million installs in its first 12 days, outpacing ChatGPT’s own mobile debut. Meanwhile, Instinct, a startup developing a personal agent you can text or call, just raised $1 billion at a $10 billion valuation. SpaceXAI introduced its Grok Bot agent in August, and most recently, on September 29, OpenAI released Dots, agents built to pursue goals in the background with minimal oversight.

So, what are personal agents? A simple way to understand them is to think of them as employees that are always on. You give them a goal, you give them the access and tools they need to achieve that goal, and then you get out of the way and let them get to work. The key new piece of the puzzle, the thing that is driving the hype, is that these goals are real-world goals—things like shopping, making dinner reservations, replying to emails, filing insurance claims, and so on.

The potential productivity gains are enormous. But there’s no such thing as a free lunch, and the upside comes with serious risks that business leaders must urgently address. This piece tells you what you need to do.

{"blockType":"mv-promo-block","data":{"imageDesktopUrl":"https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/creator-faisalhoque.png","imageMobileUrl":"https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/faisal-hoque.png","eyebrow":"","headline":"Ready to thrive at the intersection of business, technology, and humanity? ","dek":"Faisal Hoque’s books, podcast, and companies give leaders the frameworks and platforms to align purpose, people, process, and tech—turning disruption into meaningful, lasting progress.","subhed":"","description":"","ctaText":"Learn More","ctaUrl":"https:\/\/faisalhoque.com","theme":{"bg":"#02263c","text":"#ffffff","eyebrow":"#9aa2aa","subhed":"#ffffff","buttonBg":"#ffffff","buttonHoverBg":"#3b3f46","buttonText":"#000000"},"imageDesktopId":91420512,"imageMobileId":91420514,"shareable":false,"slug":"","wpCssClasses":""}}
The risk-reward trade-off

The value of a personal agent is directly related to the access and permissions you give it: The more an agent knows about your relationships, schedule, and spending, the more useful it becomes. For example, it would be incredibly useful for frequent travelers to have their flights and hotels organized by agents, but this is only possible if the agent can check calendars, write messages, and make payments.

But this same access can make agents dangerous. In February, for instance (and somewhat ironically given how hard Meta has pushed Muse), Meta’s own director of AI alignment watched an agent she had connected to her inbox start deleting it, ignoring her instruction to check with her first.

The point here is not that the risk is so great that we should never use personal agents. The point is rather that the risk cannot be separated from the reward. An agent that can’t read your inbox can’t clear it, and one that can’t spend money can’t book your flight. Every permission that makes an agent more useful also widens what can go wrong.

Personal agents will come to work

The issue for enterprise is that the access people grant doesn’t stop at their personal lives. In a recent survey of knowledge workers, employees reported giving AI tools direct access to company email, cloud storage, and customer databases, and more than half of those using unapproved AI tools said they had shared internal messages and emails with them.

And leaders may be overestimating how much of this is visible and being managed—in the same survey, 90% of executives said they were confident in their visibility into AI tools, while 52% of employees admitted to using AI tools without approval. Meanwhile, a Cloud Security Alliance survey found that 82% of organizations have AI agents running in their systems that they don’t know about, and 65% had suffered an agent-related security incident in the previous year. Jim Taylor, president of the security company RSA, describes learning about a midsize global bank that believed it had no AI agents, only to find out it was running more than 4,000.

These surveys cover AI tools and agents of all kinds used at work, but they don’t isolate personal agents such as Muse and Dots; indeed, the surveys were done before these products became widely available. But the implications for personal agents and enterprise are clear and stark—people are already hooking up their personal agents to their company emails and databases. When they do, the agent’s risks become the company’s risks.

3 risks leaders need to understand

Here are three of the most important risks that come with personal agents.

1. Your data leaves the building. This risk needs no attacker. Connect a personal agent to work email, and copies of company data go to a vendor your company has no contract with, sitting outside its procurement and compliance rules. Muse users, for instance, must opt out if they don’t want their interactions used to train Meta’s models, and we can be certain that many will forget to do so when using Muse in work contexts.

2. Outsiders can give it orders. Anyone who can email your employee can try to inject their agent instructions. One early Instinct tester emailed his own inbox from a brand-new Gmail account with instructions meant for the agent. After seeing how easily it could be phished, he deleted his account. Even Meta, which built layers of defenses into Muse, says “prompt injection remains an open problem in the industry.”

3. It’s a master key. Whoever takes over an agent inherits everything it’s connected to. Last month, a researcher showed that any app or command running on a Mac could hijack Muse and the permissions its owner had granted, and steal the token that logs into the user’s Muse account. Agents can also read the inbox where security codes arrive: Instinct was observed pulling a sign-up code from a tester’s inbox to finish a restaurant booking. In the wrong hands, that ability could unlock accounts the agent was never connected to.

The first risk can largely be handled through vendor choice and settings. The other two risks can’t be removed without giving up what makes agents useful—but they can be reduced. In all three cases, the key is setting clear limits on what agents can touch and do.

4 things to do now

Each of the steps below targets one or more of the three key risks. None requires new software, and all can start this week.

1. Find every agent this week. Since leaders may see less than they think, start by looking. Ask IT to pull the list of third-party apps connected to company email and calendar accounts. At the same time, send staff a short, no-blame survey: “Which AI agents do you use, and what have you connected them to?” By the end of the week, you should have one initial list.

2. Publish one page of agent rules. This is where you set the limits. Sort agents into three lists: approved, approved with limits, and banned. “Approved with limits” might mean authorized for use with personal accounts only. Ban connecting any agent to finance systems, customer data, password managers, or the inbox where security codes arrive; those are the connections that turn an agent into a master key. Approve only vendors that let you delete your data and opt out of training. Name one person who owns the rules page, and update it every month.

3. Require a human “yes” before any action. Because outsiders can try to steer an agent, keep a person between it and anything consequential. Agents may read, sort, and draft. But a person must approve before an agent sends email outside the company, spends money, shares a file, or accepts terms. And where possible, give the agent its own email address and forward it only what it needs, instead of opening up your whole inbox.

4. Practice switching agents off. Data can outlive the connection, so test whether disconnecting actually works. Pick one agent. Revoke its account access, change any passwords it used, and cancel any payment card linked to it. Twenty-four hours later, check whether it is still acting or still holding data. Ask the vendor in writing to delete your information. Then add “disconnect all agents” to the checklist you use when employees leave the company.

Set the terms now

Personal agents will soon be as ordinary as smartphones. Indeed, the problem reminds me of the bring-your-own-device problem of the 2010s, when employees started using their personal phones for work email. But there is a critical difference here: A phone could store your company’s data—an agent can act on it.

Just as people were using their own phones at work a decade ago, your people are already using these agents. Either you set the terms on which personal agents enter your business, or you let others do it for you.

{"blockType":"mv-promo-block","data":{"imageDesktopUrl":"https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/creator-faisalhoque.png","imageMobileUrl":"https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/faisal-hoque.png","eyebrow":"","headline":"Ready to thrive at the intersection of business, technology, and humanity? ","dek":"Faisal Hoque’s books, podcast, and companies give leaders the frameworks and platforms to align purpose, people, process, and tech—turning disruption into meaningful, lasting progress.","subhed":"","description":"","ctaText":"Learn More","ctaUrl":"https:\/\/faisalhoque.com","theme":{"bg":"#02263c","text":"#ffffff","eyebrow":"#9aa2aa","subhed":"#ffffff","buttonBg":"#ffffff","buttonHoverBg":"#3b3f46","buttonText":"#000000"},"imageDesktopId":91420512,"imageMobileId":91420514,"shareable":false,"slug":"","wpCssClasses":""}}

Job seekers are using ‘résumé Botox’ to seem younger. They say it’s helped them land interviews

1 month ago

Where cosmetic Botox softens wrinkles, so-called “résumé Botox” does the same for job seekers’ applications, making them appear younger to prospective employers. That might mean removing or editing graduation years, early jobs, senior titles, and profile photos, all in the service of obscuring an applicant’s real age.

The practice is more common than you might think. A new survey suggests that nearly half of job seekers are manipulating their résumés to come across as younger—and many are using artificial intelligence to do it.

The survey, which was conducted by Use.AI, polled 4,263 users of the AI platform who had applied for a job in the past 12 months. More than 4 in 10 (42%) said they’d removed or changed details on their résumés or LinkedIn profiles to mask their age or seem more junior.

Attempts to appear younger were more common among women than men, with 53% of women 45 and older having made these changes to their résumé, compared with 35% of men in the same age range.

These job seekers said they’re making these changes due to a fear of ageism in the hiring process, citing both human hiring managers and AI agents as possible culprits of age discrimination. Of those who had changed their résumés to appear younger, 56% said they believed human recruiters would filter out their applications, while 38% said they thought AI screeners would do the same.

Amid those AI-fueled anxieties, nearly three-quarters (72%) of those making changes to their résumés were using AI to do so. However, given that those surveyed are already users of Use.AI, their rates of AI use may not reflect those of the wider population.

Many of those who’d attempted to present themselves as younger said the strategy had paid off: 43% reported securing more job interviews after editing their résumés or LinkedIn profiles, compared with just 10% who said they’d received fewer interviews.

Ageism in the workplace

Job seekers’ fears of age discrimination aren’t unfounded. A recent survey by AARP found that two-thirds of workers 50 and older have seen or experienced ageism in the workplace, with 91% of those workers saying they believe age discrimination against older workers is common. Within that figure, 36% say they believe it is very common.

Nearly a quarter of those surveyed (22%) also reported feeling they are being pushed out of their job due to their age. 

In practice, age discrimination at work takes many forms. Many over-50 workers reported experiencing subtle expressions of age discrimination, including getting stereotyped as less tech-savvy than younger employees (33%), being seen as resistant to change (24%), having their accomplishments and expertise go ignored (20%), finding themselves the butt of jokes about generational differences (21%), and losing training opportunities to younger employees (20%).

Carly Roszkowski, vice president of financial resilience programming at AARP, said the survey results are relevant to all workers, as ageism is “the one discrimination that we all will go through.”

“We are part of an environment and a society that is obsessed with youth,” Roszkowski said. “We need to be really trying to break down those barriers and those stereotypes across all generations.”

Healthcare has a churn problem

1 month ago

I was diagnosed with anorexia at 10 years old. Over the next several years, my family and I navigated a healthcare system that was mostly focused on how many days I stayed in a hospital bed instead of whether I was getting better.

I spent my entire freshman year of high school in and out of facilities, counting down 45-minute therapy slots while fighting a brain disorder that forces you to battle your own mind upwards of six times a day. The bills came in, claims were processed, and the checkboxes were ticked, but I was still struggling with anorexia.

Decades later, now as a founder and CEO of an eating disorder treatment program, I realize my teenage experience embodied the traditional healthcare business model to a T.

THE ONLY INDUSTRY WHERE PRODUCT FAILURE IS PROFITABLE

In traditional healthcare, revenue is tied to volume: More time in session results in more revenue. The more appointments booked, the more the facility can bill; the longer someone stays in treatment—and the more times they relapse—the more profitable the facility becomes.

If any consumer tech or hardware startup ran on this model, it would fail fast. It’s like going to an auto repair shop that charges by the hour, sends you home with a car that still stalls, and then charges you again when you have to tow it back into the shop.

The main problem is industry alignment. Fee-for-service models reward visit volume and usually lead to fragmented, episodic care. Value-based outcome models, on the other hand, tie reimbursement directly to clinical and financial outcomes. If we want a healthcare system that can keep up with rising costs and demand, we need to align incentives around real, measurable outcomes.

BUILD WHAT WORKS

Tying financial compensation to real-world outcomes gives leaders the flexibility to build better products that result in positive outcomes for all parties involved. In healthcare, that means patients, their families, providers, and health insurance plans.

For example, when we started Equip, we knew that once-a-week therapy wasn’t enough to address a complex eating disorder, but the traditional fee-for-service billing system made it nearly impossible to bill for anything else. By partnering with forward-thinking health insurance plans to develop value-based payment structures, we designed the treatment model families needed: a team of doctors, dietitians, therapists, and coaches delivered virtually. This approach provides the right dose of care when patients need it, without uprooting their daily routines or stopping them from building a life worth living.

WHAT DOES “GOOD” LOOK LIKE?

The catch, however, is that the healthcare industry has historically lacked a consensus on what “good” means. Defining quality is particularly difficult in eating disorders, where recovery isn’t as simple as a blood test or a healed bone. The field uses two tools—the EDE-Q and the PARDI-AR—to set severity thresholds, but neither gets into what “getting better” means. It also requires a long-term view, which can clash with an industry where most health insurance plans operate on 12- to 24-month budget cycles.

When I was 10, that meant staying in a hospital bed long enough for my vitals to stabilize. But as I found out through many cycles of treatment, that didn’t lead to me getting better.

To solve the national mental health crisis, we need an economic framework where doing good and doing good business are the same. By replacing the traditional model with outcome-based partnerships, we aren’t just building more efficient companies, but delivering on healthcare’s biggest promise: building a product that works.

Kristina Saffran is cofounder and CEO of Equip.

AI fear is driving employees to slack off, show up late, and even steal coworkers’ lunches

1 month ago

As more businesses adopt artificial intelligence, some employees—who are increasingly worried about losing their jobs—are starting to act out.

That’s the finding of a new study from Washington State University researchers, published in the journal Occupational Health Science, that looks at how AI adoption at work is affecting workers’ attitudes and behaviors in the office.

Research has shown that the increase of AI-based technologies at work—such as intelligent robotics, automated systems, and algorithmic management—may make employees feel insecure about their jobs and lead to more stress and less engagement.

Washington State University’s study looked at employees across 18 different industries and found that workers who had the greatest fears about being replaced by AI were more likely to behave poorly by, for example, slacking off at the office or even stealing a coworker’s lunch.

Researchers also found that employees who assumed they would be displaced were more likely to “morally disengage” at work, such as by coming in late for a shift or surfing the web while on the clock.

“Moral disengagement is basically a way of justifying bad behavior,” Tahira M. Probst, a professor of psychology at WSU Vancouver and lead author of the study, said in a news release. “We know what’s right and wrong. We know what organizational policies say we can and can’t do. Moral disengagement is a way of reframing that behavior as something that is acceptable.”

Researchers used the AI Industry Exposure index to identify which industries were most exposed to potential AI disruption. Workers in finance and insurance had some of the highest levels of exposure, while workers in agriculture, forestry, and construction had the least. Not surprisingly, the bad behavior was more pronounced in the industries with the highest amount of disruption.

The study’s bottom line: When employees feel their jobs are at risk, they may look at workplace expectations differently and rationalize behaviors such as cutting corners or undermining coworkers to protect themselves or compensate for what they perceive as injustice.

The findings also suggest that AI anxieties can hurt workers and businesses alike, pointing to the need for employers to be more mindful of how they roll out AI in the workplace so employees are better supported.

Nine out of 10 CEOs plan on hiring more despite AI boom

1 month ago

Despite widespread adoption of artificial intelligence, 89% of CEOs of major U.S. corporations expect to increase overall headcount over the next three years. 

This is according to a survey released today by accounting firm KPMG, which contacted the CEOs of 400 major U.S. companies with at least $500 million in annual revenue.

The results indicate that CEOs increasingly view AI as something to integrate into their processes, rather than a way to cut labor costs. About eight in 10 said AI has led their companies to reconsider how they develop employees, while three-quarters said they have reassessed the skills needed for entry-level positions. 

Another 86% said they are “redesigning roles, career paths, and team structure to reflect human-AI collaboration.” A similar percentage (83%) said they are focused on retaining or retraining “high-potential talent,” while 79% are hiring new talent with AI and tech capabilities. 

Meanwhile, just 53% said they are planning AI-related workforce reductions in some areas. 

While CEOs expect to continue pouring money into AI—84% list it as a top investment priority over the next three years—the returns on these investments are still materializing. While roughly 40% of CEOs said AI agents had made “significant” improvements to their company’s performance, only 9% described the current impact of agentic AI as “transformational.” 

Moreover, AI adoption costs are a widespread concern. 

Roughly half of CEOs described managing AI-related costs as “mostly or very challenging,” while 81% anticipate that AI will have a greater impact on their margins than labor costs over the next five years.

More than 70% said a lack of “reliable AI access and advancement” was the biggest disruption risk facing their companies. 

As for their broader outlook, the CEOs surveyed were generally optimistic. A clear majority expressed confidence in their company (84%), industry (82%), and the broader US economy (79%). 

CEOs’ views on the global economy were somewhat gloomier, with only 64% expressing a positive outlook. Nearly 80% also said inflationary pressures will continue to affect their company over the next three years.

Southwest Airlines is introducing unlimited Companion Pass trips as part of a major loyalty overhaul

1 month ago

Your next ticket on Southwest Airlines might come with an unexpected perk: a free ticket for a friend to join you.

During a press event Thursday in Texas, the airline announced a revamp of its loyalty program, including a new Platinum loyalty tier that will give customers unlimited Companion Passes to have a friend or family member join them for their travels.

Those free flights also won’t be restricted to just Platinum flyers. The brand is also introducing new Milestone Rewards that will let customers unlock new benefits as they work toward Platinum status, including up to three Companion Pass trips before they get there.

The changes are part of a sweeping overhaul of not only the brand’s loyalty program, as Southwest transitions away from its legacy as a low-cost airline and positions itself to better compete with the likes of United Airlines, Delta Air Lines, and American Airlines.

A new way to earn unlimited Companion Pass trips

For longtime Southwest flyers, Southwest’s Companion Pass isn’t new. The perk has been one of the airline’s most recognizable loyalty benefits for a while now.  Traditionally, qualifying members could designate someone to accompany them on flights, paying only applicable taxes and fees for the companion’s ticket.

Under the new program, Southwest’s loyalty program looks a bit more like its competition. 

The airline plans to introduce three elite status levels: Silver, Gold, and Platinum. The airline’s existing A-List status will become Silver, while A-List Preferred will become Gold. Platinum will sit above both and include unlimited Companion Pass trips.

[Image: Southwest Airlines]

While unlimited Companion Pass trips are restricted to Platinum customers, travelers who don’t reach Platinum can earn new Milestone Rewards that will get them up to three passes as they move up in status throughout the year.

Southwest says the changes are intended to reward members throughout their loyalty journey rather than requiring them to reach a new status tier before receiving additional benefits.

“Rapid Rewards was built on a simple idea: make it easy for Customers to earn and use rewards,” said Nandika Suri, vice president of Rapid Rewards at Southwest Airlines, in a statement.

“As our Members engage more with Southwest, they’ve told us they want Rapid Rewards to reward them in more meaningful ways, whether they’re working toward status or using the points they’ve earned. These updates build on that feedback by rewarding Members throughout their Southwest experience, while keeping the simplicity and value that have always set our program apart.”

The unlimited Companion Pass benefit is notable because most U.S. airlines offer more restrictive companion-ticket benefits, often requiring the airline’s premium credit card.

For example, Delta Air Lines offers an annual Companion Certificate to eligible Delta SkyMiles Platinum and Reserve American Express cardholders. The certificate is issued after the card renews and can be used for one qualifying round-trip companion ticket.

Southwest is also changing how customers can qualify for its loyalty tiers.

Starting January 1, 2027, every Rapid Rewards credit cardholder will earn one Tier Qualifying Point for every dollar spent on purchases using an eligible Southwest Chase credit card.

That includes holders of the Southwest Rapid Rewards Plus card, giving customers another way to make progress toward status without relying exclusively on flights.

For frequent travelers who also put substantial everyday spending on their airline credit cards, the change could make elite status easier to achieve. Southwest will continue allowing members to qualify through flights or Tier Qualifying Points. Rapid Rewards points will also continue to have no expiration dates, and award flights will remain available without blackout dates.

Rapid Rewards points will soon work on international airlines

Southwest is also expanding where travelers can use their rewards.

Starting in the spring of next year, members will be able to earn and redeem Rapid Rewards points through new international airline loyalty partnerships.

The initial partners will be Icelandair, China Airlines, and Etihad Airways, giving Southwest customers access to destinations including Reykjavik, Paris, Taipei, Tokyo, Abu Dhabi, and Johannesburg.

The change is significant for an airline that has historically focused primarily on domestic travel, with international service concentrated in Mexico, Central America, and the Caribbean.

Members will be able to use points for seats and rental cars

Starting later this year, Southwest flyers will also have more options for redeeming their points for things other than air travel. 

Beginning in November, members will be able to redeem points for rental cars through participating rental car companies on Southwest’s booking website.

Later in 2027, they’ll be able to use points to select seats on Southwest-operated flights, including Preferred and Extra Legroom seats.

That change is particularly noteworthy given Southwest’s broader transformation.

Starlink WiFi is coming to every Southwest plane

Onboard internet has recently become a focus for several U.S.-based airlines, and Southwest is also making a major investment in onboard connectivity.

The airline plans to equip its entire fleet with Starlink satellite internet by fall 2027, offering high-speed WiFi from takeoff to landing. Despite United also investing in the service, Southwest says it expects to become the first major U.S. airline to complete a fleetwide Starlink rollout. 

The service will be free for Rapid Rewards members through the airline’s partnership with T-Mobile.

What the changes mean for Southwest travelers

The Rapid Rewards overhaul is the latest step in Southwest’s transformation from a low-cost airline known for its relatively simple policies into a carrier offering more of the premium experiences traditionally associated with its competitors.

But how accessible those benefits will be depends on the requirements for earning them. With Platinum status requiring substantial flying or credit card spending, the most valuable rewards will remain geared toward Southwest’s most engaged customers.

Florida can now enforce a ban on social media accounts for kids

1 month ago

The state of Florida won a key court decision Wednesday that allows it to enforce a ban on social media accounts for kids under 14 and other restrictions.

The law was one of the most restrictive in the U.S. on social media use by children when Gov. Ron DeSantis signed it in 2024.

The law implicates the First Amendment, an appeals court said, but “likely does so in a content-neutral manner.”

It would likely survive a full legal challenge “because it doesn’t burden substantially more speech than necessary to advance the important interest in protecting minors from social media addiction,” a panel from the 11th U.S. Circuit Court of Appeals said in a 3-0 opinion.

The law bans social media accounts for Florida children under 14 and requires parental consent for use by kids who are 14 or 15.

Industry trade groups the Computer & Communications Industry Association and NetChoice sued Florida, arguing that the law violated the free-speech rights of social media providers and users. A judge in 2025 said the law likely was unconstitutional and granted an injunction.

That injunction was put on hold while the appeals court heard the case. The lawsuit, with the injunction now overturned, will return to federal court in Tallahassee.

The appeals court noted that the law covers social media platforms “that have been determined to induce heavy use by children, and applies only to platforms that use both algorithms and one of several addictive features.”

An email seeking comment from the Florida attorney general’s office was not immediately answered. NetChoice said it was disappointed with the decision.

“We are currently evaluating all of our options as we continue the fight to ensure Floridians’ rights to free expression online are protected, and parents remain in the driver’s seat when it comes to making decisions for their families,” said Paul Taske, a lawyer at NetChoice.

—Ed White, Associated Press

Why does the internet hate Jessica Tarlov so much?

1 month ago

Jessica Tarlov is a Democratic political strategist and cohost of The Five on Fox News. As the show’s lone liberal voice in a four-to-one political dynamic, she uses the platform to represent Americans dissatisfied with the direction of U.S. politics and the Trump administration. Her goal is to model the kind of political discourse happening across the country—and show that even when common ground feels out of reach, we can still listen to, engage with, and tolerate differing perspectives.

Where’s the beef? As prices soar, midterm candidates trade barbs over meat

1 month ago

In the midterm elections, meat is on the menu.

As Democrats and Republicans argue over high prices and the Iran war, they’re also having a food fight over who’s the most red-blooded carnivore.

The latest example came this week in Mississippi, where Democratic Senate candidate Scott Colom released a video with a snippet of Republican Sen. Cindy Hyde-Smith saying that if consumers are upset over high beef prices, “you know you have so many proteins to choose from.”

The video cuts to a man grilling burgers who says, “Pick another protein? I’d rather pick another senator instead.”

The Mississippi race is somewhat unusual because it’s the Democrat hammering the Republican over beef. Elsewhere, Republicans have branded Democrats as vegans or otherwise opposed to meat, trying to show that they’re culturally out of step with voters, hostile to local cattle industries or just not manly enough.

In Texas, Republicans seized on an old interview in which Democratic Senate nominee James Talarico boasted that his campaign, then for the Texas statehouse, was going vegan. Talarico has since been publicly scarfing down barbecue to rebut allegations that his dietary choices mean he’s not a real Texan.

In Colorado, Republicans have gone after Democratic congressional candidate Manny Rutinel for his activism against meat production and factory farming. In college, he was photographed shirtless with a banner for the People for the Ethical Treatment of Animals around his chest. In 2021, Rutinel founded Climate Refarm, which advocated for cutting down on the use of meat and has signed onto letters advocating for raising taxes on meat and dairy. He also supported a 2024 ballot measure to ban a lamb slaughterhouse in Denver, which failed.

Rutinel, 31, is now running against Republican Rep. Gabe Evans in Colorado’s 8th district, which includes a vast swath of the northeastern part of the state dotted with ranches, dairy farms and meat processing facilities, including the headquarters of beef production titan JBS Swift. Republicans have hammered
Rutinel over advocating veganism, to the point that the conservative media group Daily Wire this week released an undercover video of a campaign official saying Rutinel eats meat “as little as possible.”

Rutinel, who’s made a point of publicly eating steak tacos on the campaign trail, scoffed at such arguments. “I don’t think Coloradans care about what’s on my plate, I think they care more about what’s on theirs,” Rutinel said in an interview after a recent campaign stop.

Evans said the issue goes beyond Rutinel’s meal preferences. “I don’t care what he eats personally,” Evans said in an interview. “I care that he wants to use the power of the government to make everyone eat that way.”

Republicans have gone after two other Democratic House candidates on meat — hammering Lindsay James in Iowa’s 2nd district, an area filled with hog farms, for supporting “meatless May” in a 2020 Facebook post, and Amish Shah in Arizona’s 1st district for his longstanding vegetarianism and animal rights activism.

Also in Iowa, Republican Senate candidate and current Rep. Ashley Hinson slammed her opponent, state lawmaker Josh Turek, for opposing a bill banning lab-grown meat from being sold in the state. “He’d rather have meat that comes from a lab than from our family farms,” she said at a campaign appearance.

Democrats have hammered on meat as well, but mainly just as one of the staples that cost more under Trump, squeezing voters’ pocketbooks. Trump has tried to lower prices by opening U.S. markets to beef imports, especially from Argentina, which is governed by a close Trump ally.

However, those moves have angered U.S. producers as prices have remained stubbornly high, creating a potential weakness in Republican-leaning states where Democrats hope to gain ground this year. According to the Bureau of Labor Statistics, the average pound of ground beef cost $6.79 last month, up from $5.55 in January of 2025, when Trump returned to office.

That was the dynamic that Hyde-Smith, a cattle rancher and former state agriculture commissioner, was addressing in her interview that was clipped in Colom’s video. The prosecutor — who also delivers the “pick another senator” line in the spot — is running a longshot challenge in Mississippi.

A spokesman for Hyde-Smith’s campaign argued that the Democrats deserve blame for high prices. “It is Scott Colom who supports Biden-era policies that drove up inflation,” said Nathan Calvert.

Associated Press writer Hannah Fingerhut contributed to this report.

—Nicholas Riccardi, Associated Press

Why Silicon Valley’s AI super PACs aren’t talking about AI

1 month ago

Welcome to AI Decoded, Fast Company‘s weekly newsletter that breaks down the most important news in the world of AI. I’m Mark Sullivan, a senior writer at Fast Company, covering emerging tech, AI, and tech policy.

Sign up to receive this newsletter every week via email here. And if you have comments on this issue and/or ideas for future ones, drop me a line at sullivan@fastcompany.com, and follow me on X @thesullivan.

AI Super PACs are paying millions for midterms ads, which almost never mention AI 

Political action committees backed by artificial intelligence developers, Silicon Valley executives, and venture capital firms have poured tens of millions of dollars into competitive 2026 midterm elections. But across the thousands of television, streaming, and digital ads bought with that money, the words “artificial intelligence,” “algorithms,” and “tech policy” appear in almost none of them, a New York Times investigation found.

Silicon Valley donors and corporate super PACs want to elect members of Congress who are likely to support industry-friendly AI regulation, or no regulation at all. But they must do so without further elevating AI as a campaign issue. The technology, and the data centers that support it, remain deeply unpopular with voters. Many Americans worry that AI will eliminate jobs, pose potentially existential risks, enable mass surveillance and media manipulation, and ultimately benefit only the wealthy elites who finance and control it. Wide majorities also oppose the construction of noisy, energy-hungry data centers in their communities. With AI policy now a political liability, advertisers are channeling their funds into conventional campaign ads focused on kitchen-table economic issues like inflation, local taxes, and small business support.

And we’re talking about unprecedented sums of money. A New York Times analysis of campaign finance data shows that political groups aligned with major AI developers like OpenAI and Anthropic, along with tech-focused venture capital firms, have directed between $55.7 million and $60 million into federal midterm campaign advertising and voter outreach. According to Public Citizen’s campaign finance tracking, the sector’s flagship super PAC, Leading the Future, raised $75.1 million this cycle, roughly two-thirds of it from the venture capital firm Andreessen Horowitz. Last year, OpenAI’s cofounder and president, Greg Brockman, and his wife, Anna, contributed $50 million in personal funds, splitting the money between Leading the Future and the primary Trump-aligned super PAC, MAGA Inc.

Silicon Valley’s super PAC money also doesn’t buy as much advertising as ordinary campaign cash. Federal Communications Commission rules require broadcast stations to charge super PACs undiscounted market rates, while granting candidate campaigns statutory “lowest unit rate” discounts. Because outside groups can pay up to five times as much per ad slot as candidate campaigns, tech-backed PACs are concentrating their media buys on broad, high-resonance economic messaging rather than risking millions of dollars on a topic that triggers immediate public resistance.

Under the Trump administration’s arrangement with Big AI, AI developers and infrastructure companies shower the president with cash and praise, while Trump does everything in his power to keep AI regulation voluntary, nonbinding, or nonexistent. Late last month, for example, the administration brought tech titans to Washington for a luncheon and ceremonial signing of a voluntary safety pact asking advanced AI labs to follow safety protocols and establish internal teams to monitor their own progress.

Anthropic says AI worms could soon spread across smart gadgets

Anthropic published research on Monday suggesting that small AI models may be just months away from being able to replicate themselves across cheap, internet-connected devices without relying on data centers. Researchers warn that these models could spread across devices such as smart ovens and off-brand Android TV boxes, making them difficult to shut down because there would be no central point of control. In Anthropic’s tests, GLM-5.2, a large language model developed by the Chinese AI company Z.ai, and other powerful models successfully replicated themselves across devices. Researchers predict that smaller models could achieve the same capability by the end of the year.

Huang says Nvidia will announce $2 billion in European deals

Nvidia’s CEO, Jensen Huang, said on Tuesday the company expects to unveil $2 billion worth of European AI deals in the next week, on top of the $8 billion it has deployed across the continent over the past year, Bloomberg reported. Speaking in Paris—first stop on a tour that also takes him to London and Berlin—Huang announced investments in four French startups, according to Quartz. Nvidia is already a major investor and chip supplier for the French AI company Mistral. “Europe needs to get in the race,” he said, adding that open-source models may be Europe’s best route to catch up with U.S. developers. Nvidia, of course, would benefit from a more intense global AI arms race regardless of who wins, because all players need Nvidia chips.

Hundreds of scientists call for an international AI containment body

More than 400 scientists signed an open letter calling for an international body to standardize containment practices for cyber-capable AI models and establish protocols for tracking, documenting, and reporting containment failures. The letter urges organizations developing, deploying, and testing such models to immediately adopt high-assurance containment systems, default to airgapping, and report containment failures to the proposed body; those incidents would eventually be made public. As of Wednesday, the letter’s website listed 406 signatories, including Geoffrey Hinton, Yoshua Bengio, and Stuart Russell.

Google releases Gemini 4 Argon

Google released its latest frontier model, Gemini 4 Argon, on Tuesday. According to Google, Argon ranks first on the Vals Index, which measures performance across finance, coding, legal, and tax tasks. The company says the model helped its software engineers increase agentic code submissions by more than 35% while reducing code rollbacks by nearly a third. Google also says Argon helped the Chrome team complete an architectural refactor spanning hundreds of files, compressing months of work into weeks. Argon is available globally through Google’s application programming interface (API), as well as through the Gemini app and AI Mode in Search for Ultra and Pro subscribers.

Chinese companies bought more than 1 million Nvidia gaming GPUs in a year

Chinese tech companies have purchased more than 1 million Nvidia gaming and workstation graphics processing units (GPUs) to make up for shortfalls in AI computing power, the Financial Times reported. China’s domestic AI chip production has doubled this year but still cannot keep pace with demand from the country’s largest tech companies, leaving a shortfall of nearly 1.5 million chips against demand exceeding 5 million. The stopgap chips are primarily Nvidia’s RTX Pro 6000D, RTX Pro 5000, and RTX 5090, all of which comply with U.S. export controls. While they cannot match the performance of Huawei’s Ascend 910C, they are useful for simpler AI workloads.

More AI coverage from Fast Company:

Want exclusive reporting and trend analysis on technology, business innovation, future of work, and design? Sign up for Fast Company today.

Checked
11 minutes 34 seconds ago
Fast Company
Fast Company inspires a new breed of innovative and creative thought leaders who are actively inventing the future of business.
Subscribe to Fast Company feed