What happened
Bloomberg reports that senior Trump administration officials met virtually with Anthropic’s chief compute officer Tom Brown on Sept 15 to discuss AI safety risks. The meeting involved Pentagon chief technology officer Emil Michael and Commerce Secretary Howard Lutnick. This engagement occurred amid heightened tensions following CEO Dario Amodei’s Sept 12 essay calling for AI regulation and pacing, which President Trump dismissed as a 'hoax.' The meeting takes place while Anthropic remains blacklisted by the Pentagon and is engaged in an active lawsuit against the department.
Bloomberg reports that senior Trump administration officials met virtually with Anthropic’s chief compute officer Tom Brown on Sept 15 to discuss AI safety risks. The participants included Pentagon chief technology officer Emil Michael and Commerce Secretary Howard Lutnick. A Department of Defense official stated that the department 'maintains contact with frontier AI companies' but that its stance on Anthropic has not changed.
The meeting occurred in the context of heightened tensions following a 3,800-word essay posted by Anthropic CEO Dario Amodei on Sept 12. Amodei called for government regulation and urged pacing the development of advanced AI systems to better understand potential threats. This essay was embraced by OpenAI CEO Sam Altman and SpaceXAI CEO Elon Musk, prompting President Donald Trump to dismiss AI risk fears as a 'hoax' and reject new rules.
Anthropic’s relationship with the US government has been strained since early 2026, following a disagreement with the Commerce Department over the security of its Mythos and Fable AI models and a conflict with the Pentagon over military use of its technology. The Pentagon blacklisted Anthropic, declaring it a threat to the US supply chain, and the company is currently engaged in an active lawsuit against the department in a Washington court.
Tom Brown, who has become Anthropic’s primary face in Washington, previously led negotiations to lift export controls on the company’s most advanced models. He spoke publicly with Lutnick at a G20 summit in September, where Lutnick suggested the company had 'gotten religion' and patched its relationship with the government, although the Pentagon maintained its blacklist. The Sept 15 meeting represents a continued, albeit fraught, engagement on safety issues.
The meeting coincided with broader industry discussions on AI safety. OpenAI’s global policy chief Chris Lehane stated that OpenAI has been engaging with Anthropic and Google DeepMind on safety for several weeks. At the Salesforce Dreamforce conference and an AI gathering in Washington, industry leaders and political figures, including Senator Bernie Sanders and former Trump adviser Steve Bannon, discussed the risks of advanced AI, with some warning that companies are jeopardizing human interests.
Source details: straitstimes.com ↗
Why it matters
This meeting signals a direct, high-level channel of communication between the US government and a major AI developer on safety, despite public political opposition to regulation. It highlights the complex, multi-agency nature of AI governance in the US, where the Pentagon and Commerce Department are engaging with Anthropic on safety and supply chain issues even as the President rejects formal guardrails. The interaction is significant because it occurs while Anthropic is still blacklisted by the Pentagon and suing the department, indicating that operational and safety discussions are proceeding in parallel with legal and political conflicts. It also reflects a broader industry trend, with OpenAI and Google DeepMind also engaging in safety discussions, suggesting a shift toward collaborative risk management among frontier labs and government entities.
The meeting indicates that despite public political rhetoric against AI regulation, the US government is actively engaging with frontier AI developers on safety concerns. This suggests a pragmatic approach to managing AI risks that operates alongside, and sometimes in tension with, the administration’s public stance.
The involvement of both the Pentagon and the Commerce Department highlights the multi-agency nature of AI governance in the US. The Pentagon’s focus on military AI adoption and supply chain security, combined with the Commerce Department’s role in export controls and trade, creates a complex regulatory landscape for AI companies.
The meeting occurs while Anthropic is still blacklisted by the Pentagon and suing the department. This parallel engagement on safety and legal disputes suggests that the government is attempting to manage the relationship with Anthropic on multiple fronts, potentially seeking to mitigate risks while maintaining leverage in ongoing conflicts.
The broader industry context, with OpenAI and Google DeepMind also engaging in safety discussions, suggests a shift toward collaborative risk management among frontier labs. This could lead to the development of industry-wide safety standards or protocols, potentially influencing future regulatory frameworks.
The political dimension of the meeting is significant, as it occurs amid a public dispute between President Trump and AI leaders over the need for regulation. This tension could impact legislative efforts in Congress and the broader political discourse on AI safety.
What to watch next
Watch for any public statements from the Pentagon or Commerce Department regarding the outcome of the meeting, potential changes to Anthropic's blacklist status, or new regulatory frameworks. Monitor the progress of Anthropic's lawsuit against the Pentagon. Observe whether other AI companies, such as OpenAI and Google DeepMind, engage in similar high-level safety discussions with US officials. Track the political fallout from President Trump's dismissal of AI safety concerns and how it impacts legislative efforts in Congress.
Monitor for any public statements from the Pentagon or Commerce Department regarding the outcome of the Sept 15 meeting, including any changes to Anthropic’s blacklist status or new safety protocols.
Track the progress of Anthropic’s lawsuit against the Pentagon, as the outcome could have significant implications for the company’s relationship with the US government and its ability to operate in the military sector.
Observe whether other AI companies, such as OpenAI and Google DeepMind, engage in similar high-level safety discussions with US officials, which could indicate a broader trend toward government-industry collaboration on AI safety.
Watch for legislative developments in Congress, as the political tension between the administration and AI leaders over regulation could influence the pace and scope of AI safety legislation.
Assess the impact of President Trump’s public dismissal of AI safety concerns on the broader political discourse and public perception of AI risks, as this could affect the political feasibility of regulatory measures.