The Trump administration is rolling back export restrictions on advanced "Anthropic" models.
The US administration has lifted the requirement for Anthropic to obtain a prior license to export its artificial intelligence models Mythos and Fable outside the United States, in a move that paves the way for restoring access to these models, which are among the most advanced in the world, for international users. The company announced that it will begin reactivating access to the two models starting today, July 1.
Restrictions disrupted access to the models
On June 12, the US government added the Mythos and Fable models to its list of technologies subject to export restrictions, preventing them from being made available to foreign users without special approval, according to a report published by TechCrunch and reviewed by Al Arabiya Business. Anthropic explained that compliance with these requirements was virtually impossible on a large scale, which led it to suspend public access to the two models.
New agreement with the Ministry of Commerce
After weeks of negotiations, U.S. Commerce Secretary Howard Lutnick announced that the company had agreed to a number of commitments, most notably: - Proactively detecting and addressing security risks associated with models. - Cooperation with the US government to develop protocols and standards for the release of Mythos and Fable models and future releases.
Reporting any malicious activity or misuse detected to the US authorities.
Despite announcing these commitments, Anthropic indicated that it had voluntarily pledged to implement most of these measures before the export restrictions were imposed, which led a number of cybersecurity experts to question the feasibility of the original decision. They believe that the restrictions were not primarily intended to enhance security, but may have been used as leverage against the company, especially after some Anthropic officials publicly criticized the way President Donald Trump’s administration dealt with artificial intelligence and the potential for the technology to be exploited for political purposes.
Increasing Asian competition
Anthropic had made the Mythos model available on a limited basis since April to a number of organizations, with the aim of evaluating its advanced capabilities for detecting and exploiting software vulnerabilities. In June, the company launched the Fable model to the public, adding stricter security controls to limit abuse. But as Asian companies like Fugu and Tulongfeng develop competing models with similar capabilities, the US government has faced increasing pressure to relax restrictions on Anthropic so that the United States does not lose its leadership in the global artificial intelligence race.
Uncertainty surrounds Washington's policy towards artificial intelligence
Days earlier, the Commerce Secretary authorized the release of the Mythos model to a select group of clients who had received White House approval. Similarly, OpenAI's latest models were made available to a limited number of organizations approved by the Trump administration, rather than being released to the public directly.
These developments point to the continued uncertainty surrounding the US administration’s policies regarding the regulation of artificial intelligence models, as companies still lack a clear vision of the rules that will govern the launch of future models. An executive order issued by the administration in June, which mandates a review of certain AI models before their release, has also come under fire from prominent analysts, including Dean W. Paul, who recently took up a position on OpenAI's policy team.
0 Comments