The Trump White House Is Over Anthropic’s Dario Amodei

TL;DR

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

The Trump White House has stopped direct talks with Anthropic CEO Dario Amodei, now engaging with cofounder Tom Brown. This change comes as discussions on AI model export controls continue, but key decisions remain pending.

The Trump White House has shifted its direct engagement with Anthropic, now primarily communicating with cofounder Tom Brown rather than CEO Dario Amodei, according to sources familiar with the matter. This change occurs amid ongoing discussions about lifting export controls on Anthropic’s AI models, which are currently restricted due to security concerns. The move signals a shift in the administration’s approach to AI regulation and its relationship with the company, which has implications for future AI policy and export decisions.

Sources indicate that the Trump administration has found it more productive to speak directly with Tom Brown and Anthropic’s public policy chief, Sarah Heck, rather than with Dario Amodei, who previously represented the company in negotiations. One person familiar with the talks described Brown as more cooperative and better able to engage with the administration’s concerns.

The administration has not yet lifted the export controls that have kept Anthropic’s most powerful models offline since June 12. These restrictions were imposed after the NSA flagged potential security risks related to the models’ guardrails and capabilities. Discussions are ongoing about what proof or assurances Anthropic might provide to alleviate these concerns and enable the model’s re-deployment.

Meanwhile, a bipartisan group of lawmakers has raised questions about the criteria and timeline for lifting these export restrictions. They sent a letter to Commerce Secretary Howard Lutnick, asking for clarity on the decision-making process and expected timeline, with responses due by June 26. The Commerce Department has not yet confirmed whether it will meet this deadline or provide further details.

Implications of Leadership Change in AI Negotiations

This shift in communication from Dario Amodei to Tom Brown reflects internal dynamics within Anthropic and may influence the pace and direction of negotiations with the U.S. government. The administration’s preference for engaging with Brown suggests a desire for a more straightforward or cooperative dialogue, which could impact the future of AI export policies and the company’s ability to deploy its models internationally. The ongoing restrictions and discussions highlight the complex balance between fostering AI innovation and addressing national security concerns.

Background on AI Export Controls and Anthropic’s Role

Anthropic, founded by former OpenAI researchers including Dario Amodei, has developed advanced AI models such as Fable 5 and Mythos. The U.S. government imposed export controls on these models in June, citing national security risks related to their capabilities and potential jailbreaks. These restrictions aim to prevent misuse but have also limited the company’s ability to operate freely abroad. Learn more about export controls.

Prior to this shift, Dario Amodei was the primary representative in negotiations, but sources say he was difficult to work with and did not adequately address the administration’s concerns. The current engagement with Brown and Heck marks a strategic change, possibly reflecting a different approach to resolving security issues and regulatory hurdles.

“Tom Brown is not being a weirdo like Dario and can actually engage.”

— an anonymous source familiar with the negotiations

Unclear Timeline and Conditions for Model Re-Deployment

It remains uncertain when or if the export restrictions will be lifted, as negotiations about proof and safeguards continue. The specific criteria the government will require from Anthropic are not yet publicly known, and the timeline for a decision remains unclear. The outcome depends on whether Anthropic can satisfy security concerns related to jailbreak risks and guardrail robustness.

Next Steps in AI Export Policy and Company Negotiations

The key next step is the response from the Commerce Department regarding the lawmakers’ questions and the potential lifting of export controls. Further negotiations between Anthropic and government officials are expected to focus on security assurances and technical safeguards. The timeline for a final decision remains uncertain, but developments could occur within the coming weeks.

Key Questions

Why did the White House stop talking directly to Dario Amodei?

Sources say the administration found Tom Brown more cooperative and better able to engage on security concerns, leading to a shift in communication focus.

What are the main security concerns about Anthropic’s models?

Officials are worried about jailbreaks and the potential to disable guardrails, which could enable misuse of powerful AI models.

When might the export restrictions be lifted?

The timeline is uncertain; it depends on whether Anthropic can provide sufficient proof to address security concerns. No specific date has been announced.

What role do lawmakers play in this process?

Lawmakers have asked the Commerce Department for clarity on criteria and timelines for lifting restrictions, influencing the regulatory process.

Could this shift affect Anthropic’s ability to deploy models internationally?

Yes, if restrictions are lifted, it could enable broader deployment, but current negotiations focus on security assurances before any change occurs.

Source: WIRED