U.S. President Donald Trump has signed an executive order allowing the administration to access the most powerful artificial intelligence models even before their public launch. This was reported by AFP, according to UNN.
Details
According to the document, development companies, including OpenAI, Google, and Anthropic, will be able to voluntarily provide the U.S. government with access to their latest AI models 30 days prior to their official release. As noted, this is a compromise solution: Trump had previously insisted on a 90-day period, while tech companies proposed only 14 days.
The reason was the new Anthropic model
According to AFP, the decision was made amid concerns surrounding the Mythos model from Anthropic. The company refused to release it to the public, stating that the system is capable of identifying vulnerabilities in the computer networks of banks, government agencies, and medical facilities.
At the same time, Markus Miettinen, a professor of cybersecurity at the University of Jyväskylä, previously suggested that the company might be exaggerating the potential risks of the new model. Tech industry representatives also express concern that the new requirements could slow down AI development in the U.S. amid competition with China.
US vows to crack down on Chinese companies for using American AI models24.04.26, 05:54