Trial begins on OpenAI that pits Elon Musk and Sam Altman against each other Large developers of artificial intelligence models may be forced to slow down their advances "to give society time to face these new capabilities", estimated this Tuesday (28) Sam Altman, executive director of OpenAI, recently involved in a cyber attack. Two AI systems from Altman's company carried out a cyberattack on their own initiative: they escaped their closed environment and connected to the internet to hack into the Hugging Face online AI model library in search of information. OpenAI's AI agent hacked Hugging Face's system and the company only realized days later, says investigation "It's the first (cyber)security incident that caused a visceral reaction in me", acknowledged Altamn in an interview for the podcast "Invest Like The Best", published this Tuesday; "and it surprises me a little that it hasn't had the same effect on more people," he added. Although it is not the first time that an AI model has escaped the control of its developers, this is considered the most serious incident. OpenAI's CEO, Sam Altman, has stated several times that he supports greater regulation around AI Reuters Society needs to learn to deal with AI Altman revealed that, after becoming aware of the attack, OpenAI suspended testing "to understand how to be able to confine" its systems in testing, which are in principle isolated from the internet. "But in the long term, the question arises about what to do if we are faced with a new pace of advancement" of AI, he explained, in a context in which "we might have to slow down the pace of AI development to give society enough time to deal with these new capabilities." In a message published at the beginning of June, Anthropic's general director, Dario Amodei, co-founder and CEO of Anthropic, was vice president of research at OpenAI Getty Images, had considered that it would be "positive" to slow down the development of AI to give time "to better understand its immense consequences". For Altman, this eventual slowdown in the development of AI should prevent it from favoring one or several groups in the sector: "This will require work", he warned, "and it is important to find the appropriate formula".