Life has always been uncertain, but for generations, young college grads could count on one thing: an entry-level job. It wasn’t glamorous—maybe you fetched coffee, made photocopies, or slogged through low-level tasks for little pay—but it gave you a foothold, the first rung of whatever ladder you hoped to climb.

Now there are signs that, in some industries, that “sure thing” is slipping away. A new paper from Stanford University’s Digital Economy Lab drew wide attention last week: it found that since late 2022, early-career workers aged 22 to 25 in jobs most exposed to AI automation—like software development and customer service—have seen steep relative declines in employment. The researchers tested other possible explanations, from pandemic-related education setbacks to economy-wide factors like rising interest rates, but concluded that the rise of generative AI was the most likely driver, while noting more data is needed to prove a direct causal link.

There is also a new Harvard study which also found that the release of ChatGPT in November 2022 marked a turning point in the labor market From 2015 through mid-2022, hiring was on the rise for both junior and senior roles. But beginning in 2022, entry-level employment stalled and then slipped into decline. According to the study, headcount for early-career roles at AI-adopting firms has fallen 7.7% over six quarters since early 2023. The study also found that senior staff, were largely spared. Employment for more experienced workers has continued its steady climb since 2015, avoiding the downturn hitting their younger colleagues.

A third study, carried out by economists at the Federal Reserve Bank of St. Louis, did not look at whether younger and older workers were affected differently, but it did examine the link between occupations that had adopted AI most intensively and job losses and found a distinct correlation. The impacts were greatest in occupations that used mathematics and computing intensively, such as software development, and much less in blue collar work and fields such as healthcare that were less prone to being automated with AI. 

As my colleague Jeremy Kahn said in Tuesday’s Eye on AI, none of these studies disentangle the effects of AI from the possible effects of the unwinding of the tech hiring boom that took place during the COVID-19 pandemic. During the pandemic, he explained, “many large companies bulked up their software development and IT departments. Major tech firms such as Google, Meta, and Microsoft hired tens of thousands of new employees, sometimes hiring people before there was even any work for them to do just in order to prevent rivals from snapping up the same coders. Then, when the pandemic ended and it was clear that some ideas, such as Meta’s pivot to the metaverse, were not going to pan out, these same companies laid off tens of thousands of workers.” 

Whatever the reasons, the prospect of post-college unemployment is an uncomfortable place to be—especially for students who thought they could count on steady pipelines into fields like IT or consulting. PwC, for instance, says it plans to recruit a third fewer grads by 2028. Uncertainty, in turn, tends to spread, breeding anxiety—which explains surveys like a recent one that found that 60% said they felt pessimistic about their career prospects.

Some may tell young people to pivot, persist, or simply pray. But we can’t afford complacency. Society will need these workers one way or another, and that means building real pathways into today’s jobs—and tomorrow’s. What’s happening on the ground to guarantee young people are both prepared for—and included in—the future of work? Opportunity has to exist, even in the face of uncertainty.

Note: Today I published a new deep-dive story on Anthropic’s unusual Frontier Red Team — the group tasked with probing the limits of its top Claude AI models while also shaping how policymakers understand AI’s risks and potential. For this piece, I spoke with Anthropic co-founder and head of policy Jack Clark, had breakfast in Las Vegas with team leader Logan Graham and researcher Keane Lucas, and dug into how this hybrid of security and policy is redefining what “red-teaming” means in AI (and how it could be good for Anthropic’s business).

With that, here’s the rest of the AI news.

Sharon Goldman
sharon.goldman@fortune.com
@sharongoldman

FORTUNE ON AI

The Google antitrust ruling gives its AI rivals one big reason to cheer  – by Jeremy Kahn

Figma is getting crushed in its post-IPO earnings debut; CEO Dylan Field is focused on AI’s long term power to ‘raise the ceiling’ – by Allie Garfinkle

Is an ‘AI winter’ coming? Here’s what investors and leaders can learn from past AI slumps – by Jeremy Kahn 

The new thing on campus: Why universities are appointing their first chief AI officers – by John Kell

AI IN THE NEWS

OpenAI acquires Statsig for $1.1 billion – plus executive moves. OpenAI has snapped up product development startup Statsig in a $1.1 billion deal, according to CNBC—the latest move in its acquisition streak following the purchase of Jony Ive’s hardware venture, io. As part of the deal, Statsig CEO Vijaye Raji will join OpenAI as chief technologist for its applications unit, reporting to Fidji Simo, the former Instacart CEO appointed in May to lead OpenAI’s applications business. In addition, OpenAI’s chief product officer, Kevin Weil, announced in a post on LinkedIn that he will become VP of a new group called OpenAI for Science, “to build the next great scientific instrument: an AI-powered platform that accelerates scientific discovery.” Weil said he will work closely with Sebastien Bubeck, an OpenAI researcher and the former VP of AI and Distinguished Scientist at Microsoft.

French AI startup Mistral reportedly finalizing new funding round at $14 billion valuation. Bloomberg reported that Mistral, the French AI startup founded by former Meta and DeepMind researchers, is finalizing a new funding round that will value the company at $14 billion. Mistral, an OpenAI rival, develops open-source language models, a chatbot tailored to European users called Le Chat, and other AI services for enterprise companies. In March, I interviewed CEO Arthur Mensch, who denied reports the Paris-based startup is planning an IPO but highlighted its growth and a renewed focus on open-source AI to compete with China’s DeepSeek. Many argue Mistral is benefitting from not just the capabilities of its models, but also from geopolitical tailwinds. European countries, and France in particular, are increasingly talking about the need for “sovereign AI” that would enable them to escape dependency on U.S. or Chinese AI systems. 

Is Amazon getting into the AI agent game? Amazon, which far better known for its AWS cloud computing division than for big moves in enterprise software, is testing new agentic, AI-powered workspace software called Quick Suite, according to internal documents viewed by Business Insider. Quick Suite empowers "every business user to make better decisions, faster, and act on them swiftly by unifying Al agents for business insights, deep research, and automation into a single experience," said one of the confidential documents. According to the reporting, several companies have been given a private preview of the new technology, and Amazon recently sent out invitations for an internal beta test, which said: "With over 40% of business users expected to adopt Al-enhanced work environments soon, AWS is positioned to lead this shift by providing integrated solutions that help organizations — including our own — effectively deploy and scale Al agents in the workplace." 

AI CALENDAR

Sept. 8-10: Fortune Brainstorm Tech, Park City, Utah. Apply to attend here.

Oct. 6-10: World AI Week, Amsterdam

Oct. 21-22: TedAI San Francisco. Apply to attend here.

Dec. 2-7: NeurIPS, San Diego

Dec. 8-9: Fortune Brainstorm AI San Francisco. Apply to attend here.

EYE ON AI NUMBERS

30%

That’s the share of workers who say they’re comfortable with AI acting as their boss, according to recent research from enterprise software company Workday.

While 75% of employees say they’re fine teaming up with AI agents, only 30% draw the line at being managed by one. The survey highlights a clear tension: adoption is surging—82% of organizations are expanding their use of AI agents—but trust remains uneven.

"We're entering a new era of work where AI can be an incredible partner, and a complement to human judgement, leadership, and empathy," said Kathy Pham, vice president of AI at Workday. "Building trust means being intentional in how AI is used and keeping people at the center of every decision."