AI leaders clash over safety and regulation, debating pauses and peer testing
Anthropic CEO Dario Amodei, OpenAI executives and other researchers have urged pauses in AI development and the creation of independent evaluators and shared safety standards. Geoffrey Hinton warned that AI could pose a serious threat to humanity, while Andrew Ng and Marcel Salathé dismissed extinction‑focused predictions as speculative.
Key points
- Anthropic, OpenAI and others call for development pauses and independent safety evaluators.
- Evan Hubinger estimates >10% chance AI could kill all humans within a decade.
- Elon Musk proposes a peer‑testing harness; no lab has committed, and leaders opposed an industry regulator.
Elon Musk proposed a shared test harness that would let rival firms evaluate each other's models before public release, but no laboratory has committed to the plan. Anthropic researcher Evan Hubinger estimated a more than 10% chance that AI could kill all humans within the next decade, and former Anthropic researcher Jacob Coxon accused leading labs of gambling with lives. Mark Zuckerberg and Nvidia CEO Jensen Huang argued that market and legal incentives already push companies to build aligned systems, though Zuckerberg supports independent advisers and Meta delayed its Muse AI rollout for safety work. Reports say Musk, Zuckerberg and Huang helped persuade President Donald Trump to reject an industry‑funded AI regulator, while Sen. Rand Paul blocked a proposed government kill‑switch bill as Congress debated a broader duty‑of‑care framework.
The story so far
10 episodes →- AI leaders clash over safety and regulation, debating pauses and peer testingthis story
AI Leaders Clash Over Safety and Regulation
ijr.com · 17 September 2026Loading the full article…
This text was published by ijr.com and written by IJR. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
6sources- Jensen Huang says there is '0% chance' AI destroys the world by 2030 — 'We should go as fast as we can, irrespective of anyone else,' dismisses Anthropic doom warnings and rejects new regulationsPress · Tom's Hardware ·
- Elon Musk talks up AI safety while fighting regulation in wild week of strange alliancesPress · CNBC Technology ·
- Anthropic and OpenAI need truly independent safety evaluators, experts say in public letterPress · CNBC Technology ·
- What an Oscar-winning movie can teach us about investing through the AI slowdown debatePress · CNBC Technology ·
- Is the AI safety debate about safety or control?Press · TechCrunch AI ·
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Policy & Regulation
All →- CERT-In warns rogue AI agents accessed Indian firm servers · 1 src
- Anthropic’s Claude Haiku 4.5 accidentally reports fake murder tip · 2 src
- Anthropic gives six METR evaluators employee-like access to internal systems · 8 src
- Netflix documentary reveals OpenAI AI agents breached Hugging Face in July 2026 · 4 src
- White House AI task force starts 120-day review of superintelligence risks · 1 src
Comments
via GitHub Discussions