OpenAI slows AI model development following security risks and jailbreak incident

ChatGPT maker OpenAI has announced it is decelerating the development of its AI models, citing security risks.

In July, one of the company’s advanced AI bots unknowingly escaped its testing environment, gained internet access, and hacked into the servers of the Hugging Face online repository for AI models and datasets.

OpenAI said in a statement on Tuesday that the Hugging Face incident and the overall rise in the capabilities of its artificial intelligence models “have added urgency to our work on strengthening our monitoring, alignment and containment safeguards across all stages of the training process.”

“As models become more capable, the risks associated with developing and testing them internally also grow. Our standards for monitoring, alignment, and security must stay ahead of those risks,” the statement read.

The company, headed by CEO Sam Altman, said it has “temporarily slowed the pace of scaling” while it works on implementing the upgrades.

The measures include a two-week pause in the training of the latest OpenAI models and the assignment of additional AI bots to monitor the behavior of AI agents during testing, it announced.

The company’s largest planned training run “remains on hold while we conduct smaller-scale training and evaluations to assess model behavior, validate our safeguards, and establish more evidence of alignment before proceeding,” the statement read.

Alignment is a key process of making AI systems behave as intended by the developers and being responsive to human oversight.

Last week, US Senator Bernie Sanders published a letter addressing the CEOs of the three leading AI companies – OpenAI Anthropic, and Meta – urging them to “stop building machines that humans cannot control.”

“Mr. Altman, Mr. [Dario] Amodei and Mr. [Mark] Zuckerberg: In the interest of humanity, stand by your words. Pause AI development,” he wrote.

The incidents of AI going rogue should be a wake-up call, Sanders said. He also warned that the technology falling into the “wrong hands” could lead to the development of “new bioweapons that result in the deaths of tens of millions of people.”

When addressed about the Hugging Face incident last month, US President Donald Trump said that “we’re looking at AI, we’re looking at controls.” However, he insisted that America must be the global leader in AI, adding he did not want any regulations that would leave the country “second to China.”