More than 200 economists, AI researchers, and Nobel laureates signed a statement today titled 'We Must Act Now,' warning that AI could drive an economic transformation larger than the Industrial Revolution over the next 10 years. Signatories include executives from Anthropic, Google, and OpenAI, alongside MIT's Daron Acemoglu and Simon Johnson, the pair who won the 2024 Nobel in economics and have historically been AI skeptics. The statement calls on policymakers to build 'incentives, guardrails, and institutions' to steer AI toward complementing human workers rather than replacing them.
The letter itself is short and includes no specific policy ask. Its weight comes from who signed it. Erik Brynjolfsson, the Stanford economist who organized the effort, told the New York Times the signatures represent a shift in economic thinking, with researchers now taking large-scale disruption more seriously than they did even a year ago.
The statement lands a month after a Wall Street Journal survey of 16 leading economists found half expect AI to produce no net change in jobs, five expect net losses, and three expect net growth. Last week, China opted not to set a numerical target for urban job creation over the next five years — the first time it has skipped that target since the 1990s.
Key facts
- 01More than 200 economists, AI researchers, and Nobel laureates signed the 'We Must Act Now' statement, including executives from Anthropic, Google, and OpenAI.
- 02Signatories include Daron Acemoglu and Simon Johnson, the MIT professors who won the 2024 Nobel in economics and have previously been AI skeptics.
- 03Stanford's Canaries Dashboard shows early-career jobs shrank 2.7% this year while mid-career jobs (ages 35-40) grew 1.6%.
- 04US software development job postings are up 15% since Claude Code launched in February 2025, but 71% of that increase is for senior-level roles.
- 05A Wall Street Journal survey of 16 economists last month found half expect no net job change from AI, five expect losses, and three expect growth.
The data underneath the warning is mixed but directionally consistent. The Yale Budget Lab reported in June 2025 that overall US employment shows no clear AI signal yet.
But zoom in and the picture shifts. Stanford's Canaries Dashboard, another Brynjolfsson project drawing on ADP payroll data, found this month that jobs most exposed to AI shrank 0.5%, while the least exposed grew 0.2%. Early-career jobs shrank 2.7% this year, while jobs for mid-career workers aged 35 to 40 grew 1.6%. The effect sizes are small, but the pattern points at entry-level work.
That reframes what looked like a happy headline last week. US job postings related to software development are up 15% since Claude Code launched in February 2025 — but 71% of that increase is for senior-level roles. Employers appear to be using AI to absorb the junior-engineer workload while still hiring seniors to review, direct, and debug the output. The obvious question is what happens when the models get good enough to handle the senior tasks too.
Economists disagree about how much of any of this is actually AI. Interest-rate hikes, pandemic-era overhiring, and remote-work aftershocks all continue to drag on hiring. And AI's contribution to job displacement is hard to reconcile with the fact that measurable productivity gains from the technology remain thin. Corporations also have obvious incentives to attribute layoffs to AI when the real cause is a bad quarter.
The policy debate is starting to catch up. There is now bipartisan support in the US for a sovereign wealth fund funded by AI companies. Labor economist Kathryn Anne Edwards has proposed overhauling unemployment insurance and funding worker relocations. The Brookings Institution's Molly Kinder has floated wage insurance and government incentives for employers to hire younger workers. None of these has serious legislative momentum yet.
Brynjolfsson told the Times his top ask is better data. Different measures currently tell contradictory stories about which workers are being affected and by how much, and the lack of reliable, granular employment data has made it hard for researchers to say anything definitive. Without that data, policy debates default to anecdote and vibes.
The counterweight worth naming: prior waves of automation produced predictions of mass unemployment that did not arrive on the timelines forecasters expected. Platformer's own miniseries on AI and jobs, which spoke to seven experts, ended mostly in optimism, with most guests casting doubt on long-term mass unemployment while acknowledging most jobs will change. The 200-signatory letter is not a prediction of catastrophe — it's a request that governments build the institutional capacity to respond if one arrives.
For AI companies, the letter is a strategically useful document. Signing it lets Anthropic, Google, and OpenAI say they take the labor consequences of their products seriously without committing to any specific product change or pricing concession. The real test comes when a US administration proposes an actual AI-funded transfer program, retraining mandate, or hiring incentive. The signatures on 'We Must Act Now' will be worth watching against the lobbying disclosures that follow.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
