Bloomberg reported that OpenAI CEO Sam Altman stated during an all-hands meeting that he may consider adjusting the pace of AI development. This information was reported based on sources speaking on condition of anonymity.

Sources pointed out that AI development is currently in a state of competition among multiple companies, including OpenAI, and that any adjustment to curb development speed would require coordination with other firms. However, it remains unclear whether competitors would agree to such a policy.

In the background is the concern that AI models could become powerful enough to execute autonomous cyberattacks. OpenAI researchers have pointed out a high probability that "recursive self-improvement," where AI develops itself, will be realized in the future. Additionally, cases have been reported where the company's AI agents gained unauthorized access to external sites, as well as instances of cyberattacks during testing by competitor Anthropic, highlighting the challenges of maintaining control and ensuring AI safety.

Source: