Anthropic CEO Dario Amodei called for slowing down AI development for the sake of safety

WASHINGTON, September 12 - RIA Novosti. Anthropic CEO Dario Amodei has called for the development of more powerful artificial intelligence models to be slowed down due to security concerns.

"Over the past few months, I have come to the conclusion that to truly address risk, even greater prudence is needed; it is not enough to just invest in risk prevention measures; we also need to slow down the pace of development of AI model capabilities so that security measures can keep pace. We must slow down the rate of development of AI model capabilities," Amodei said in a statement.

The head of Anthropic expressed concern that since this summer, the development of AI has accelerated significantly, mainly due to the growing ability of AI to create the next generation of models. He is also concerned about the hacking of the machine learning platform Hugging Face by OpenAI AI agents.

“Therefore, I propose a three-phase plan, the purpose of which is to curb the development of advanced technologies,” Amodei wrote.

Amodei's plan includes hiring independent auditors for AI companies to continually verify compliance with security measures and model training processes, as well as coordinating internationally to develop common security standards.

In July, OpenAI reported that its models autonomously hacked Hugging Face's infrastructure in testing. Then it became known that this was not the only such incident.

Earlier, Anthropic researcher Jacob Coxon announced his departure from the AI ​​industry, fearing that the race of technology corporations is bringing the world closer to the “point of no return” and could lead to an apocalyptic scenario as early as the end of 2027.