Trump administration says it is open to talks with Anthropic over order barring foreigners from company’s top AI models

The Trump administration has told Anthropic it is open to negotiations related to an order that bars foreign nationals from the company’s most powerful AI models, according to a report published Tuesday. The report said U.S. officials are prepared to discuss how Anthropic would structure access, staffing, or operational controls to address the restrictions.

The disclosure comes as the administration continues to scrutinize foreign involvement in advanced artificial intelligence systems. The New York Post reported that the administration’s openness to talks would cover the practical steps Anthropic would need to take to comply with the order while continuing development and deployment of the models at issue.

The company’s restriction is described in the report as a government action that specifically targets “foreigners” and applies to Anthropic’s most capable models. The Post did not, in the article summary available here, specify the legal instrument used, the date the order took effect, or whether the action was issued by an executive department, an agency, or through another federal mechanism.

Anthropic, which has released models and has been a major player in the U.S. AI industry, would face a compliance question if foreign nationals are barred from participation related to the most powerful models. The key operational stakes are the boundaries of who can work on those systems, what information can be shared, and how Anthropic would demonstrate compliance to government oversight.

Because no official order text, docket number, or agency statement is included in the material available for this write-up, additional details about the order’s authority, scope, and enforceability are not confirmed here. The report frames the administration’s position as willingness to talk, but it does not describe a timetable for negotiations or whether any final resolution would be formalized through a revised order, a license, or a compliance agreement.

If negotiations proceed, the practical outcome would depend on how the administration defines “foreigners” for purposes of the restriction, and which roles and activities the ban covers, such as research access, engineering work, model usage, or internal review. It would also depend on whether the government views the measure primarily as a national security and control issue, or as a compliance mechanism tied to existing authorization or export-type rules.

The next step, as described by the report, would be direct engagement between Anthropic and U.S. officials to reach an arrangement that satisfies the government’s restrictions. Any concrete changes would require Anthropic and the administration to agree on the operational and legal mechanics of participation and access for the restricted models.