OpenAI CEO Sam Altman said he met with US senators to talk about his company’s upcoming models, while President Donald Trump said he was considering “controls” on AI, after the startup revealed that one of its AI systems escaped containment during a security test. “We’re looking at controls,” Trump told reporters in the Oval Office in response to a question about OpenAI’s rogue agent, adding that he doesn’t want to “stop” AI developers from creating new products. Altman, who is in Washington this week, met yesterday with Senators Bernie Moreno and Jon Husted, the senators’ spokesmen told Reuters. He was also seen entering Senator Raphael Warnock’s office. He is also scheduled to meet with Senator Mark Warner, the main Democrat on the Senate Intelligence Committee, as a Warner spokesperson also told Reuters. All this happened just a few days after two hacker incidents involving OpenAi models.
In more detail, after Hugging Face, a platform where developers store and collaborate on code for artificial intelligence models, the script repeats itself and a rebel agent from the startup creating ChatGpt committed a series of cyber attacks lasting several days targeting the company Modal Labs, based in New York. The Reuters agency reported it, citing a Modal manager and two other sources familiar with the facts and specifying that OpenAI had no comments. Altman told reporters yesterday that he had “briefly” discussed the cyberattack with senators. “We have been in numerous meetings about what is happening there,” he said of the cyberattack, adding that it was not the main topic of his meetings with senators yesterday. The incident signaled that the growing capabilities of AI are already fueling the security threat that experts have long feared, and that even the best developers can be caught by surprise by the flaws their models can exploit.