The Bosses of the Biggest AI Companies Say: Time to Slow Down
Dario Amodei, who leads Anthropic, and Sam Altman, who leads OpenAI, agree that AI development needs to slow down.
That agreement was reported on September 15, 2026 Associated Press. Both men run leading AI labs. Both labs are still releasing new models and business tools.
Amodei spoke more directly. He asked AI companies to "slow the rate at which they advance model capabilities" Reuters. He then wrote an essay calling for slowing down AI development, according to September 14 reporting. He also called for rules requiring "independent audits" of the AI industry.
Political leaders reacted in different ways. U.S. President Trump rejected warnings about AI safety in September 2026. He said the U.S. "already has tools to police the AI industry" Reuters. A state newspaper in China said Anthropic's call to slow down was a "Cold War tactic" Reuters.
OpenAI has published its own safety warnings. One official page on its website is titled 'An Alien Mind.' It says this moment calls for "extreme caution." It also says "no one is prepared for the consequences of a continued rapid rise in machines."
Another official OpenAI page is titled 'AI progress and recommendations.' It is dated November 6, 2025. It says OpenAI is "deeply committed to safety."
These warnings sit next to many new product announcements. On September 16, 2026, OpenAI published 'Reimagining advertising with AI,' 'How to connect AI usage to business value' and 'Our framework for reporting model misalignment.'
More releases came the week before. On September 11, OpenAI published 'Rapidly scaling online storage to serve over 1 billion ChatGPT users.' On September 10, it published 'How a researcher uses Codex and ChatGPT to search for new antimicrobial molecules,' 'Now everyone can put data to work,' 'Introducing ChatGPT for Financial Services,' 'Build more natural voice experiences with GPT-Live-1 in the API' and 'Introducing the Agents API.'
Anthropic is also releasing new products. On September 1, 2026, it announced Claude Fable 5.1 and Claude Mythos 5.1. The company calls them its "most advanced models for coding and knowledge work."
The broader context here is the difference between making the AI itself stronger and making the tools around it better. It is like building a stronger engine versus improving the rest of the car. Making the model stronger means training on more data, adding new problem-solving skills, letting it act with less supervision, and dealing with gaps in testing. Improving products means better packaging, faster voice replies, connecting tools together, storing data, and following rules for fields like finance. OpenAI's September releases are mostly in that second group. Amodei's audit idea deals with the first group, with outside checks, repeated tests and shared reports. Trump's view is that current rules are enough. Beijing treats the slowdown call as politics, not a technical issue.
In my view, the size of these systems explains the concern. A service with over 1 billion ChatGPT users, also used to search for new medicines, is no longer a small experiment. Reports about mistakes, outside audits and data storage are everyday needs at that size. My kids grew up with search, then phones, then AI chat. Each change felt slow until schools and workplaces had to catch up fast.
The reason for hope here is that careful rules help people actually use the technology. Clear reports about mistakes, outside audits, better proof of business value and safety rules for banking let powerful tools be used in daily life. If labs slow down the core models but compete on safety and reliability, working professionals get tools they can trust. In that way, caution helps adoption.
