Trump halts key regulation for AI models
A new regulation was intended to enable critical review of new AI models. However, the tech industry raised concerns with the US president at the last minute.
(Image: ChiccoDodiFC/Shutterstock.com)
A new regulation that would have allowed the US government to review new AI models before their release has been put on hold for the time being. US President Donald Trump was actually scheduled to sign the new regulation from his office on Thursday, but postponed it indefinitely. He told the press that “certain aspects” of it had not pleased him.
He believes the new regulation could interfere with the US's global leading position in AI development, Trump explained, pointing to the lead over China in the field. He therefore wants to avoid any step in such a direction.
Initially, Trump and his administration had concerns in other areas: the regulation was intended as a response to potential cybersecurity risks posed by new, advanced AI models, such as Anthropic's Claude, which the manufacturer itself deems too dangerous to make publicly available. Federal agencies would have had up to 90 days before release to scrutinize the models, for example, the domestic intelligence agency National Security Agency, the White House Cyber Office, or the US Department of the Treasury. A similar agreement already exists with AI manufacturers Google, Microsoft, and xAI.
Tech CEOs dissuaded Trump; Sacks lectured him
It is striking how Trump's sudden change of course apparently came about. Tech investor David Sacks reportedly influenced Trump shortly before the regulation was to be signed, according to a report by Politico magazine. Several individuals close to Trump, whose identities Politico does not wish to disclose for source protection, reportedly told the magazine about a conversation between Sacks and Trump. Sacks is said to have convinced Trump that AI companies were already cooperating sufficiently with the US government, that the planned regulation would hinder innovation, and that it would slow down the US in the race for the best AI. According to Politico's sources, Sacks was not the only opponent of the new regulation among the leaders of major AI manufacturers.
Videos by heise
According to another Politico report, key industry figures such as Meta CEO Mark Zuckerberg, OpenAI CEO Sam Altman, and Anthropic CEO Dario Amodei were aware of the planned regulation long before. A meeting with Trump was even planned shortly before the regulation was to be signed. But the three tech CEOs flatly stood the president up. As Politico learned from informed circles, all three canceled the meeting but wanted to send representatives. Only Sacks remained, and Trump was apparently thoroughly lectured by him, leading him to back down from the regulation he himself had commissioned.
(nen)
