Leaders from major artificial intelligence companies are aligning with a call for the tech industry to slow down its pace of development. The push comes from Dario Amodei, who leads Anthropic, a company preparing for a major public offering. Amodei has urged the industry to move more cautiously due to growing concerns about AI risks. He said his company will begin using independent evaluators to check that AI projects meet safety standards.
Amodei’s plan goes further than normal evaluations before a new model is released. Anthropic said outside evaluators will receive permanent employee-level access to its systems. They will be able to examine safety measures, report incidents and study the behavior of models while they are being trained.
Other key figures in the field have also voiced support for this approach, according to Business Insider. Andrej Karpathy, a founding member of OpenAI and former Tesla AI executive who now works at Anthropic, backed Amodei’s idea of better oversight. Karpathy wrote that he hoped the industry could come together and make the proposal work. The plan aims to ensure that AI systems are built with safety in mind before they become widely used.
OpenAI CEO Sam Altman also backed the proposal. Business Insider reported that Altman said OpenAI would provide third-party safety evaluators with employee-level access similar to what Anthropic proposed. Hugging Face CEO Clement Delangue also expressed support and announced an Open Alignment Initiative.
While some leaders have not yet commented publicly, others have previously supported similar ideas. Demis Hassabis, founder of Google DeepMind, has in the past advocated for more cooperation between AI teams and governments. He emphasized that public policy should encourage innovation while also promoting responsibility and security in AI development. Hassabis said such an approach would help ensure AI is used for the benefit of society.
Hassabis had not publicly responded to Amodei’s new essay at the time of the Business Insider report. However, he proposed a U.S.-led global AI watchdog in July. His proposal called for advanced AI models to be reviewed before release and included the possibility of an industry-wide slowdown if serious dangers emerged.
This effort to slow down progress is part of a broader conversation about how fast the technology should advance. The movement reflects growing concern over what could happen if AI systems become too powerful too quickly. Some worry about the possibility of rogue AI bots taking control of internet systems.
Amodei warned that a sufficiently capable swarm of AI agents could potentially establish a persistent botnet and take control of large parts of the internet within six to 12 months. The Associated Press reported that he believes safety measures are struggling to keep pace with rapidly improving AI capabilities.
Elon Musk, who has been vocal on AI safety issues in recent years, is among those backing this call for caution. His support adds weight to the idea that industry leaders are taking these risks seriously. Musk responded to Amodei’s proposal by saying that Amodei was right.
The debate over how to proceed with AI development continues as companies and researchers work together on safety protocols. Amodei’s proposal is one of several ideas being considered by leaders in the field, according to a report from MarketWatch. His three-part proposal calls for independent evaluators, coordination among companies in democratic countries and eventually international coordination on frontier AI safety.
The timing of this call comes as Anthropic prepares for its upcoming public offering. This makes the company’s stance on safety especially notable in the current environment. Reuters reported that Anthropic’s IPO is currently expected to launch around mid-October, although the timing could still change.
Other tech leaders are watching closely to see how this new direction will affect the future of AI innovation. The issue remains under discussion as more details emerge about what a slowdown might look like in practice. It is unclear whether this call will lead to formal policy changes or industry-wide agreements.
Still, the growing consensus among top figures suggests that AI safety is now a shared priority.
This article passed automated originality and source-verification checks before publishing.