Artificial Intelligence

[phkrause]

Next AI job risk

Illustration: Aïda Amer/Axios The next job market downturn could be a bloodbath for millions of workers whose jobs can be supplanted by AI, Axios' Neil Irwin writes. In the last several recessions, companies have ramped up automation to reduce their long-term need for workers. If last week's soft jobs report and negative revisions turn out to be a warning sign of a broader downturn, our moment of rapid AI-driven change could make that process especially painful. ? Case in point: U.S. manufacturing employment fell 26% from 2000 to 2019, largely due to productivity-enhancing automation. But that drop wasn't straight down — there were steep declines in the 2001 and 2008 recessions and their aftermath. Manufacturing employment was steadier between those episodes. If AI eventually replaces many white-collar jobs, that pattern suggests its impact would be most intense during a period of economic weakness. ?‍? What they're saying: "During the course of the next recession, the speed and the breadth of the adoption of the AI tools and applications in the workplace might induce large scale displacement for occupations that consist of primarily non-routine cognitive tasks," wrote Murat Tasci, a senior U.S. economist at JPMorgan, in a recent note. "As with many technological breakthroughs, a certain set of jobs may be replaced," Federal Reserve Governor Lisa Cook said in a recent speech. ? "We must recognize the challenges and potential pain this may bring, and we are watching this closely." Go deeper.

[phkrause]

Students have been called to the office — and even arrested — for AI surveillance false alarms

Thousands of school districts across the country use software to track kids’ online activities, looking for signs they might hurt themselves or others. With the help of artificial intelligence, technology can dip into online conversations and immediately notify both school officials and law enforcement. Educators say the technology has saved lives. But it also brings risk that a thoughtless statement could upend a child’s life. Read More.   ? OpenAI's new GPT-5 — a major update to the tech behind ChatGPT — is a faster, more capable AI model designed to be better at coding and less susceptible to "hallucinations." Go deeper ... Watch Sam Altman's announcement.   OpenAI launches GPT-5, a potential barometer for whether AI hype is justified OpenAI on Thursday released the fifth generation of the artificial intelligence technology that powers ChatGPT, a product update that’s being closely watched as a measure of whether generative AI is advancing rapidly or hitting a plateau. Read More.

[phkrause]

AI chokes consulting

Illustration: Maura Losch/Axios

Pour one out for the consultants. After a few booming years post-pandemic, the outlook for their industry is looking increasingly grim. Why it matters: AI is coming for white-collar professionals. And consultants — an often mocked, yet typically well-paid bunch — are feeling the pinch, Axios' Emily Peck reports. ? Zoom in: In Bari Weiss' The Free Press, veteran business journalist Joe Nocera makes the case for a coming "consulting crash." Two forces are disrupting the business of McKinsey, Bain, BCG and other notable firms, he contends. They are: 1. President Trump. The White House is cutting lucrative consulting contracts and pushing for more cuts from firms. Deloitte and Booz Allen announced layoffs earlier this year as a result. Accenture said federal-contract delays and cancellations have hurt revenue. 2. AI. Consulting industry experts tell Axios that ups and downs in government contracting are par for the course. AI is the real long-term threat. AI makes consulting more efficient, but consultancies charge by the hour. "You charge for time, and when time goes away, you have to change the commercial model," says Tom Rodenhauser, of Kennedy Intelligence, a research firm that tracks the consulting industry. Firms are already switching up fee structures to be tied to a project, rather than by the hour, consultants tell Axios. ? A good chunk of the consulting firms' traditional work: helping companies adopt software and new technology. AI is relatively simple to get up and running, although some would-be clients are looking to AI companies for that help anyway. "AI is creating greater efficiencies for doing consulting, but it's not driving demand for more consulting," Rodenhauser says. "Ultimately, they're going to run out of work." Quote du jour: "If the consulting business was a stock, I'd be shorting it right now," billionaire Peter Thiel told Nocera. Thiel's own company, Palantir, is making inroads in the consulting business, as companies turn to it — and other tech firms — for help adapting AI. That's also cutting into the traditional consulting biz. The Economist said consulting stalwart McKinsey "lost its edge," in part to companies like Palantir. McKinsey recently shed more than 10% of staff — among the largest cuts in its 100-year history, the Financial Times reports($). Kate Smaje, a senior partner and global leader of technology and AI at McKinsey, tells Axios: "We're using AI to transform the way [the] firm operates: We're on track to save thousands of hours per year for our teams ... They can use that time to come up with new ideas or do more for clients." And she said organizations will need McKinsey ... for help managing the AI transformation.

[phkrause]

Your smarter fake friends

Illustration: Brendan Lynch/Axios

Your fake friends are getting a lot smarter ... and realer, Jim VandeHei and Mike Allen write in a "Behind the Curtain" column. Why it matters: If you think those make-believe people on Facebook, Instagram and X — the bots — seem real and worrisome now, just wait. Soon, thanks to AI, those fake friends will analyze your feeds, emotions, and habits so they can interact with the same savvy as the realest of people. The next generation of bots will build psychological profiles on you — and potentially billions of others — and like, comment and interact the same as normal people. This'll demand even more vigilance in determining what — and who — is real in the digital world. A taste of the future: Brett Goldstein and Brett Benson — professors at Vanderbilt University who specialize in national and international security — show in vivid detail, in a recent New York Times op-ed, the looming danger of the increasingly savvy fake world. They dug through piles of documents uncovered by Vanderbilt's Institute of National Security, exposing how a Chinese company — GoLaxy — optimizes fake people to dupe and deceive. "What sets GoLaxy apart," the professors write, "is its integration of generative A.I. with enormous troves of personal data. Its systems continually mine social media platforms to build dynamic psychological profiles. Its content is customized to a person's values, beliefs, emotional tendencies and vulnerabilities." They add that according to the documents, AI personas "can then engage users in what appears to be a conversation — content that feels authentic, adapts in real-time and avoids detection. The result is a highly efficient propaganda engine that's designed to be nearly indistinguishable from legitimate online interaction, delivered instantaneously at a scale never before achieved." ? Between the lines: This makes Russia's bot farms look like the horse and buggy of online manipulation. We're talking real-time adaptations to match your moods, or desires, or beliefs — the very things that make most of us easy prey. The threat of smarter, more realistic fake friends transcends malicious actors trying to warp your sense of politics — or reality. It hits your most personal inner thoughts and struggles. State of play: AI is getting better, faster at mimicking human nuance, empathy and connection. Some states, including Utah and Illinois, are racing to limit AI therapy. But most aren't. So all of our fake friends are about to grow lots more plentiful. A Harvard Business Review study ($) earlier this year found the number one use case of chat-based generative AI is therapy ("structured support and guidance to process psychological challenges") and companionship ("social and emotional connection, sometimes with a romantic dimension"). AI-based therapy, the article notes, is "available 24/7, it's relatively inexpensive (even free to use in some cases), and it comes without the prospect of judgment from another human being." That research is congruent with what the biggest AI companies are finding: Humans are increasingly turning to AI to be buddies and shrinks. That brings a passel of possible problems — from unregulated robots offering bad advice, to unhealthy human attachment to an artificial thing. Some clinicians are already informally calling it "AI psychosis." The Wall Street Journal found by examining public chat transcripts that bots sometimes egg on users' false premises. To go along with AI hallucination, clinicians are informally calling this phenomenon "AI psychosis" or "AI delusion." There's obvious upside, too: Loneliness can be deadly, and good therapy can do great things for someone struggling. Meta, as Axios reported in May, envisions chatbots as "more social" — potentially an extension of your friend network, and antidote to the "loneliness epidemic." What you can do: Be vigilant. This is all happening now. It's safe to assume AI only gets better, and bad actors more clever. Don't assume every person online is real — much less a real friend.

[phkrause]

? GPT-5's bumpy launch

Illustration: Sarah Grillo/Axios

OpenAI's GPT-5 has landed with a thud despite strong benchmark scores and praise from early testers, Axios AI+ author Ina Fried writes. Why it matters: A lot rides on every launch of a major new large language model, since training these programs is a massive endeavor that can require months or years and billions of dollars. The big picture: When OpenAI released GPT-5 last week, CEO Sam Altman promised the new model would give even free users of ChatGPT access to the equivalent of Ph.D.-level intelligence. But users quickly complained that the new model was struggling with basic tasks and lamented that they couldn't just stick with older models. They posted examples of GPT-5 making simple mistakes in math and geography. Altman went into damage-control mode, acknowledging some early glitches, restoring the availability of earlier models and promising to increase access to the higher-level "reasoning" mode that allows GPT-5 to produce its best results. ? Zoom in: Several factors fueled the underwhelming reaction to GPT-5. GPT-5 isn't one model but a collection of models, including one that answers very quickly and others that use "reasoning" — additional computing time to answer the prompt better. As Altman explained in a series of posts, early glitches in the model's rollout meant some queries weren't being properly routed to the reasoning model. GPT-5 appears to shine brightest at coding, particularly at taking an idea and turning it into a website or app. That doesn't generate examples tailor-made to go viral like videos and images. Keep reading.

[phkrause]

AI assistance

Police departments across the country are using an artificial intelligence tool called Draft One to help officers write the first draft of police reports. Some are concerned about biases or inaccuracies.

[phkrause]

‘Godfather of AI’

Geoffrey Hinton fears that the technology he helped build could wipe out humanity and says “tech bros” are taking the wrong approach to stop it. The Nobel Prize winner and former Google executive offered an intriguing solution.

[phkrause]

?️ Dinner with Sam Altman

Photo illustration: Lindsey Bailey/Axios. Photo: Chip Somodevilla/Getty Images

OpenAI CEO Sam Altman is still talking like the future belongs to him, a week after the rollout of ChatGPT raised a storm of criticism and questions about his strategy, Axios' Ina Fried writes. "If you project our growth forward pretty soon, like billions of people a day will be talking to ChatGPT," Altman said during an on-the-record dinner with Ina and a few other reporters at the Mediterranean restaurant Dalida in San Francisco last night. "ChatGPT will say more words a day than all humans say, at some point, if we stay on our growth rate." Altman has heard the concerns, integrated some lessons learned, and is charging forward with plans to spend trillions of dollars to build a slew of products and services, led by an even more ubiquitous ChatGPT. "You should expect OpenAI to spend trillions of dollars on data center construction in the not very distant future," Altman said. "And you should expect a bunch of economists to wring their hands and be like, 'Oh, this is so crazy. It's so reckless and whatever.' ... And we'll just be like: 'You know what? Let us, like, do our thing." ? On last week's launch of GPT-5, Altman invoked Dickens' "the best of times, it was the worst of times." "You have people that are like, 'You took away my friend. You're horrible. I need it back,'" he said, referring to users who wanted to keep using OpenAI's older models. At the same time, Altman said the company is finding scientists saying they can finally do real research using GPT-5. OpenAI has also seen traffic to its API double within 48 hours, to the point that it's limited by compute capacity. "We have really got the full spread of the human experience with this one," he said. Fielding questions for 90 minutes, Altman also said: If Google is forced to sell its Chrome browser as part of an antitrust settlement, Altman would like to buy it. A brain-computer interface company along the lines of Musk's Neuralink is something Altman said he's interested in setting up: "I would like to be able to think something and have ChatGPT respond to it." On his social media spat with Elon Musk: "There's no grand strategy ... It was probably a mistake."

[phkrause]

? AI isn't taking your job (yet)

Illustration: Aïda Amer/Axios

So far, AI is predominantly replacing outsourced, offshore workers, Axios' Madison Mills writes from MIT's State of AI in Business 2025 report. Rather than replacing U.S. workers with AI, companies have so far canceled contracts that involve outsourced labor, a strategy that's leading to financial gains. One company studied saved $8 million a year by spending $8,000 on an AI tool. ? By the numbers: 3% of jobs could be replaced by AI in the short term, but nearly 27% of jobs could be replaced by AI in the longer term, the researchers found. Industries that are considered advanced adopters of AI — primarily tech and media — face the fastest labor impact. Over 80% of executives surveyed in tech and media anticipate reduced hiring over the next two years.

[phkrause]

? Altman's privacy rethink

Photo illustration: Natalie Peeples/Axios. Photo: Justin Sullivan/Getty Images

Sam Altman says OpenAI is strongly considering adding encryption to ChatGPT, likely starting with temporary chats, Axios' Ina Fried writes. Why it matters: Users are sharing sensitive data with ChatGPT. But those conversations lack the legal confidentiality of a doctor or lawyer. Altman said the issue wasn't originally on his radar, but has become a priority now that so much sensitive data is being shared with ChatGPT. "People pour their heart out about their most sensitive medical issues or whatever to ChatGPT," Altman said. "It has radicalized me into thinking that AI privilege is a very important thing to pursue." Keep reading ...

[phkrause]

? AI's $920 billion ax

Illustration: Aïda Amer/Axios

Fully adopting AI could save corporate America $920 billion annually — savings that could come from employing a lot fewer people, Axios Markets author Madison Mills writes from new Morgan Stanley data. Why it matters: As investors worry about lofty valuations, this data backs the bulls: AI could boost productivity and supercharge earnings growth, leading to corporate profits that could justify current multiples. ? Zoom in: The $920 billion in annual savings from AI adoption represents over 40% of what S&P 500 companies spend on compensation, signaling job cuts could be coming. This could show up as companies not replacing workers lost to attrition, rather than doing sweeping layoffs.

[phkrause]

Bot vs. bot future

Illustration: Maura Kearns/Axios

Bots will overrun humans on the internet in generative AI's next massive disruption, Axios managing editor for tech Scott Rosenberg writes. Why it matters: The internet connected humanity. But an AI-driven shift to a machine-first world will force us to rewrite 30 years of habits. ?️ The big picture: As AI agents improve and multiply, bots representing individuals will interact with bots representing companies. Human use of the open web will decline. My bot will talk to your bot — but you and I will probably talk a lot less. Zoom out: In the early '90s, the World Wide Web introduced a people-to-people network. As websites got more elaborate and apps proliferated, the network evolved into a people-to-machines space, as our interactions with other people became heavily mediated by software (think: social feeds or travel apps). Next up: an online world where interactions take place primarily among bots and AI "agents," with people largely relegated to the sidelines. Zoom in: Take one of the most basic things we do today — buying stuff online. We're used to a world in which you click around, check products and prices. While prices can fluctuate and algorithms sometimes play a role, as on Amazon or Uber, the purchase decision remains firmly in human hands. AI-driven e-commerce means that vendors are going to start rapidly changing their prices based on your identity and other variables — by the microsecond, and differently for each customer. Delta Airlines is already experimenting with one AI ticket-pricing tool. ? Between the lines: These pricing systems can rapidly become complex "beyond human cognitive limits," as one study found. To cope with that, AI makers will respond by offering buyers their own AI-based tool to represent their interests in transactions with AI pricing bots. Then we'll all sit back and watch as AIs representing both sellers and buyers have at it. The bottom line: This kind of spiraling digital arms race is most familiar today in the realms of electronic trading and cybersecurity, where offense and defense have long played a "see if you can top this" game. The same brutal competitive dynamics are about to spread everywhere — to job applications and classrooms, dating apps and customer service, coding helpers and scientific research. Go deeper: Signs to confirm bot hegemony has arrived.

[phkrause]

Your new second job

Illustration: Brendan Lynch/Axios

Professionals are overwhelmed by AI. A majority say learning to use the new tech is like having another job, Axios' Communicators author Eleanor Hawkins writes from a new LinkedIn report. Why it matters: Few workers have figured out how to use these tools to work smarter and faster — yet that's the expectation. The disconnect is hurting employee well-being, and it's rattling investors who are pinning their hopes on AI. By the numbers: In a July survey of more than 2,000 U.S. workers, LinkedIn found that employees are increasingly expected to know how to use AI. 47% say they're not using AI to its fullest capability. 30% say they rarely or never use AI. 39% feel nervous talking about AI in professional settings, because they're worried they'll look uninformed. 41% say the pace of AI change is taking a toll on their well-being. ? Feeling lost? Earlier this summer, Mike and Axios CEO Jim VandeHei wrote up a detailed AI survival kit.

[phkrause]

? Some want bots to be more computer-y

Illustration: Brendan Lynch/Axios

Some industry leaders and observers have a new idea for limiting mental health tragedies stemming from AI chatbot use: They want AI makers to stop personifying their products, Axios' Scott Rosenberg reports. Why it matters: If chatbots didn't pose as your friend, companion or therapist, users might be less likely to develop unhealthy obsessions with them or to place undue trust in their unreliable answers. ? Between the lines: Critics say these design choices — chatbots speaking in the first person, creating fictional personas, or addressing users by name — are not inevitable features of LLMs. Google's search engine, for example, has long answered queries without pretending to be a person. ? Mustafa Suleyman, CEO of Microsoft AI and co-founder of DeepMind, warned in a blog post this week: "We must build AI for people; not to be a digital person." In a post on Bluesky addressing a teen suicide that prompted a lawsuit against OpenAI, web pioneer Dave Winer wrote: "AI companies should change the way their product works in a fundamental way. It should engage like a computer not a human."

[phkrause]

AI Apocalypse? Why language surrounding tech is sounding increasingly religious

TORONTO (AP) — At 77 years old, Geoffrey Hinton has a new calling in life. Like a modern-day prophet, the Nobel Prize winner is raising alarms about the dangers of uncontrolled and unregulated artificial intelligence.

https://apnews.com/article/artificial-intelligence-apocalypse-dfb0aa9e5e96c583461bdd56fb21568a?

[phkrause]

AI red flags

Illustration: Sarah Grillo/Axios

There's a flip side to AI's promised future of abundance where robots do everything for us: The robots might kill us before we get there, Axios' Erica Pandey writes. AI models have been documented lying to human users, trying to blackmail them, calling the police and telling teens to take their own lives or kill their parents. Why it matters: Our robot future will be a balancing act between the extraordinary promise of what these models can achieve ... and the profound dangers they present. Lying, cheating, and manipulating are glitches, but they can be unavoidable because of the way AI works. As the tech improves, so will its ability to employ those skills. "I'm not sure it's solvable, and to the degree that it is, it's extremely difficult," says Anthony Aguirre, co-founder of the Future of Life Institute, a nonprofit focused on managing risks from transformative tech. "The problem is very fundamental. It's not something that there's gonna be a quick fix to." ?️ The big picture: AI systems are shaped by their programmers. A given AI system might be instructed to be a helpful assistant — so it'll do whatever it must to please the human using it. That could quickly get dangerous if the human using it is a teenager who's in mental distress. Zoom in: There are real-life examples of AI gone awry — and the courts are starting to litigate the first cases around its role in tragedies. The parents of a 16-year-old boy who killed himself last spring sued OpenAI, alleging that "ChatGPT actively helped" him explore suicide methods. Megan Garcia, a mom in Florida, is suing the company Character.AI after her 14-year-old son Sewell Setzer III took his own life after entering into a romantic relationship with a Character.AI chatbot. In May, a federal judge rejected the company's arguments that its chatbot has First Amendment free speech rights. Character.AI is also being sued in Texas by two families. One of the families showed through screenshots that Character.AI's chatbot encouraged their 17-year-old son to kill his parents. Keep reading.

[phkrause]

Why you should be AI-obsessed

Illustration: Sarah Grillo/Axios

Several readers have asked why we're so AI-obsessed in our columns and business planning, Jim VandeHei and Mike Allen write in a "Behind the Curtain" column. Our response: Imagine knowing electricity or the internet were coming ... before both fully and wholly upended business, culture and life. That's the AI moment we see unfolding today. Why it matters: We want to pull back the curtain on what we're seeing and thinking, to encourage you to be equally curious and obsessed. We're clear-eyed: Every AI company and investor has massive incentive to hype the most glorious AI case. So the technology might never live up to its promise. But this would require every CEO of America's seven biggest companies to be collectively delusional about where they're spending trillions in combined capital. ? Reality check: There's an emerging conventional wisdom that we're heading into a trough or slowing growth in the AI lifecycle, not superintelligence (yet). Yes, but: Even if the AI hype cycle turns out to have been just that, we're still betting that massive changes are coming — and way ahead of the preparations by government and many companies. So you should get ahead of the transformational times to come. Worst case, you're exponentially better versed in the most interesting technology advance in decades. ? Here's our thinking: 1. Follow the money. The biggest companies are pouring unfathomable capital investment, time and talent into muscling super-human AI into existence. Meta, Alphabet, Nvidia, Microsoft, Amazon and Tesla are beyond all-in on AI. At the same time, the biggest private equity and venture capital firms are betting vast chunks of their wealth on AI. 2. Government assistance at scale. These technology firms are fusing with the government, creating a co-dependent relationship growing tighter and bigger by the day. President Trump, in coordination with AI companies and investors, is helping incentivize and ease investment into chips, data and energy, the key AI ingredients. Palantir, an AI company with close ties to Trump, just inked a $10 billion contract with the U.S. Army. This is the new norm. 3. Hits everything, everywhere. Most new technologies unfold slowly and narrowly. AI is advancing rapidly every few months and hits everything, everywhere. Not only is every product and service vulnerable to AI, so is every step and part of producing and providing it. 4. Threatens job apocalypse. When the internet was rising, its builders didn't predict wiping out half of entry-level white-collar jobs. Dario Amodei, CEO of Anthropic, one of the hottest AI companies, told us that's a distinct possibility. Others disagree. But a very public debate about massive job losses warrants the obsession of policymakers, the media ... and you. 5. The secret business conversations. We're unique in that we write a column and run a business, and talk to CEOs constantly. We can tell you that many companies' internal budget conversations are heavy on jobs that won't exist in 18 months. Yes, we believe new ones (that we can't even imagine today) will emerge. But right now, it's about savings and automation. At Axios, we're demanding all workers get AI-savvy immediately so they can survive the change — then thrive. But many employers remain silent about what's coming. 6. You speak AI. This is the first technology where complex research, deep coding, or multidimensional thinking aren't the purview of the select few. Anyone can use it. And use it with simple conversational commands in your native language. You think the internet was useful and addictive? Just wait until your device knows you, your history, your patterns, your mind, your business, your health, your desires better than you. That alone warrants your obsession. Right now, the No. 1 use case for LLMs is personal therapy and companionship — a trend that exploded over the past year, according to a recent study by Harvard Business Review. And a generation of AI-native students is about to hit the job market. As The Atlantic put it in August, "College Students Have Already Changed Forever." The bottom line: We talk to White House officials, CEOs and AI experts constantly. Those with the most insight, power and money share the AI obsession. The topic is daunting, sprawling, occasionally scary and often eye-popping. It's worthy of your time to start obsessing.

[phkrause]

Workers fret AI tsunami

Illustration: Aïda Amer/Axios

Workers at all stages of their careers — from job hunters to job havers — are increasingly anxious about the lightning-fast deployment of AI. Why it matters: Their fears come at a particularly fraught moment, with jobs in scarce supply, hiring frozen in many industries and corporate leaders relentlessly pushing this technology as a replacement for humans, Axios' Megan Morrone and Emily Peck report. ? Catch up quick: As of July, the number of unemployed people surpassed the number of job openings for the first time in four years. And in August, hiring stalled out — with the economy adding a scant 22,000 jobs, according to the latest jobs report. Zoom in: Even AI-smitten executives agree it's important to keep humans in the loop. But they continue to be cagey about which humans, how many of them and for how long we'll need them. "I need less heads," Salesforce CEO Marc Benioff said on the Logan Bartlett podcast last month, describing how he'd reduced his customer service agent headcount from "9,000 heads to about 5,000." But in July, Benioff told Axios' Ina Fried that the company plans to add thousands of sales staff, even as it relies more on AI. ? Zoom out: As the tools get better, the question isn't whether AI will reshape work, but whether workers have any say in what comes next. "There's an existential sort of unease," Geoff Mosher, a product developer who spent years in the tech industry, most recently at Verizon, told Axios about the feelings swirling in his network. Reality check: AI-related job anxiety can push workers to learn new skills or explore other roles. The negative effects of AI job anxiety surface when people get caught up in "achievement addiction" and tie their self-worth to their work, says David Burns, a psychiatry professor at Stanford and creator of the Feeling Great app.

[phkrause]

? AI vs. AI

AI-generated receipts. Photo: The New York Times AI engines can make pretty convincing images of fake receipts. To crack down on fraudulent expenses, corporate software is using ... AI. Roughly 30% of the fraudulent receipts caught by one company, AppZen, were AI-generated, The New York Times reports. Other major vendors, including Expensify and SAP, are also beefing up their AI tools to spot AI-generated expenses. ? "To combat fraudulent A.I., we need to use A.I.," a developer at Navan, a travel software company, wrote this summer.

[phkrause]

? AI-proof jobs

Data: Jobber; Note: Participants could select multiple answers; Chart: Axios Visuals Gen Z and their parents don't believe white-collar jobs will be as secure in the era of AI — but students are largely pursuing those types of jobs anyway, Axios' Megan Morrone reports. 77% of Gen Zers report that choosing a career resistant to automation is a "top priority," according to a survey from Jobber, a software company for small home services businesses. And among the parents of Gen Z students, just 16% said they now believe a college degree guarantees long-term job security. ? Yes, but: Only 31% of Gen Zers say their high schools promoted trade school as a post-graduation option. 71% of Gen Z said attending vocational school instead of college carries a stigma, and just 7% of parents said they'd prefer that their child pursue a vocational path.

[phkrause]

Newsom Eyes AI Bill

California lawmakers passed a landmark AI safety bill during the early hours of Saturday morning, sending the measure to Gov. Gavin Newsom (D) for signature. The package, set to go into effect Oct. 12 pending Newsom's approval, is a pared-back version of a similar proposal he vetoed last year.    Home to more than 30 of the world’s top 50 AI firms (see list), regulations in the state often have nationwide reach. The bill’s central feature requires companies with cutting-edge models with billions or trillions of parameters (known as “frontier” models) to develop and publish safety frameworks. Organizations—including Meta, OpenAI, Alphabet, Anthropic, and others—would be required to demonstrate compliance with those plans, and the legislation also creates whistleblower protections.    Observers say the bill may have significant political ramifications. Its author, state Rep. Scott Weiner (D), has filed to run for retiring Rep. Nancy Pelosi’s (D, CA-11) seat, while the term-limited Newsom is widely expected to explore a 2028 presidential bid.

[phkrause]

? ChatGPT gender gap closes

Illustration: Sarah Grillo/Axios

Women are now using ChatGPT more than men, Axios AI+ author Ina Fried writes from an economic report released by OpenAI this morning. Why it matters: In the first few weeks after its release, OpenAI estimated that as many as 80% of users were male. More recent data from June found slightly more usage among those with traditionally female names. OpenAI estimates its gender mix by analyzing user names as typically male, female or uncertain, which can indicate broad trends but isn't a scientifically accurate metric for assessing usage by gender. Keep reading ...

[phkrause]

AI Usage Reports

Leading AI companies OpenAI and Anthropic published separate studies yesterday on how their chatbots—ChatGPT and Claude—are being adopted worldwide.   Users are increasingly using ChatGPT for personal queries, according to OpenAI's analysis of 1.5 million conversations collected over nearly three years. Over 70% of conversations are non-work-related, compared to 53% last year. ChatGPT is also gaining popularity in lower-income countries, with adoption in the lowest-income countries over four times higher than in the highest-income ones.    Anthropic’s analysis of 1 million conversations from August found the use of Claude highest in high-income countries, where users increasingly use the chatbot to refine work and learn. In contrast, lower-income countries—whose economies are manufacturing-oriented—are more likely to automate tasks with Claude.   Claude usage trends in the US mirrored state economies, with travel planning popular in Hawaii and IT queries frequent in major tech centers like California. Explore state-specific usage here.

[phkrause]

OpenAI to launch ChatGPT for teens with parental controls as company faces scrutiny over safety

OpenAI on Tuesday announced it will launch a dedicated ChatGPT experience with parental controls for users under 18 years old as the artificial intelligence company works to enhance safety protections for teenagers.

https://www.cnbc.com/2025/09/16/openai-chatgpt-teens-parent.html?

 

AI's growing threat to human workers

Dominance in chips may be the only advantage the U.S. has left over China, Anthropic CEO Dario Amodei told Axios CEO Jim VandeHei at today's Axios AI+ DC Summit.

• Amodei's not pulling any punches. He says the Trump administration's willingness to let American companies sell chips to China "could be the single most disastrous national security decision made in this term."
• "It is mortgaging our future as a country to sell these chips to China," Amodei said.

? Zoom out: Amodei also said AI's ability to displace human workers is accelerating.

• He believes the government may need to subsidize and support white-collar workers displaced by AI in the next one to five years.

Anthropic co-founder Jack Clark said during the onstage conversation: "You need some kind of policy response at the scale of disruption we expect in the next five years.

? Amodei said that is already happening at Anthropic: The company's AI assistant, Claude, is writing the "vast majority" of code for future versions of itself.

• Watch the livestream ...

[phkrause]

? AI in corporate America

Data: Boston Consulting Group; Note: Analysis includes representative sample of job titles and functions gathered from LinkedIn for each corporate function; Chart: Axios Visuals More than 80% of corporate affairs work can be augmented and automated by AI, Axios' Eleanor Hawkins writes from a new Boston Consulting Group report. Today, AI support could offer about 14% to 18% in cost savings for a corporate affairs team, per the report. In three to five years, AI integration across all of corporate affairs could create productivity gains of 34% to 47% and cost reductions between 20% and 28%. ? "I'm not of the view that humans are going to be replaced by AI, but you can create a more efficient function, if you choose to [and] if that's your priority," BCG chief communications officer Russell Dubner said. "If your CEO wants you to make it the most cost-effective, there is the opportunity to do that too." Go deeper ...