Donald Trump has signed an executive order aimed at strengthening oversight of the most advanced artificial intelligence models in the United States. The text provides for the creation of a framework to evaluate the cyber capabilities of models in order to identify those deemed strategic. Companies may voluntarily participate in this program and provide the government with access to their technologies up to thirty days prior to public release. The administration will also be able to intervene in the selection of partners authorized to conduct early testing of these models.
The executive order specifies that it creates neither a mandatory licensing system, nor a requirement for prior government authorization to develop or market AI models. This initiative comes as several major industry players, including Anthropic, OpenAI, and xAI, are preparing for or considering initial public offerings that could value their businesses at several hundred billion dollars. The administration is thus seeking to increase its visibility into the most sensitive technologies without imposing a restrictive regulatory framework.
The measure follows debates sparked by Claude Mythos Preview, an Anthropic model specialized in identifying IT vulnerabilities. The order also directs the Department of Defense to strengthen the protection of its information systems. This decision is part of a context of persistent tensions between the Pentagon and Anthropic, whose technology currently remains restricted in certain national defense contracts, a decision the company is contesting in court.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.