G7 Discusses "Trusted Partners" Access to Anthropic AI

What happened at the G7 AI access talks?

G7 leaders discussed a plan to grant select "trusted partners" access to advanced U.S. AI models at the annual summit in Evian-les-Bains, France, on June 16, 2026. Three diplomatic sources confirmed the talks to Reuters. The discussions focused primarily on Anthropic's restricted models, as Reuters reported via U.S. News.

The Financial Times first reported the "trusted partners" scheme. Reuters then confirmed the details through its own diplomatic sources at the summit.

Why did Anthropic cut off access to its models?

U.S. President Donald Trump ordered Anthropic to block foreign nationals from using its most advanced AI models, citing national security concerns. In response, Anthropic disabled access to Fable 5 and Mythos 5 — its two most advanced models — for all users last Friday.

Fable 5 and Mythos 5 are Anthropic's top-tier AI models. Mythos 5 is specifically designed to find flaws in computer code. Cybersecurity experts believe Mythos could accelerate attacks on banks' technology systems.

Before Trump's order, Anthropic had given Mythos access to select organizations in more than 15 countries. Those organizations used the model to scan their computer systems for vulnerabilities. They included entities in the healthcare, communications, power, and water sectors.

The broader Anthropic ban's impact on open-source AI is already being tracked by analysts watching how the restrictions reshape the competitive landscape.

You might also like

SpaceX Eyes $20B Bond Sale Ahead of 2026 IPO

OpenAI o3 Solves 18 Pediatric Rare-Disease Cases

Amazon May Sell Trainium Chips Directly to Rivals

Microsoft Sells OpenAI Models to ByteDance, Hitting $1B

Who led the G7 AI access discussions?

One diplomatic source said country representatives discussed widening AI model access mainly with U.S. Commerce Secretary Howard Lutnick. Those talks happened on the sidelines of the opening G7 summit dinner on Monday in Evian-les-Bains.

A second source, who declined to be named because talks were ongoing, said the "trusted partners" could be either countries or companies. No final agreement was announced.

A Trump White House official stated that the president's team has "an open line of communication with our allies, and we remain committed to addressing national security concerns with Anthropic's model."

Here's what we know so far: the talks are ongoing, no deal has been confirmed, and the structure of any future access arrangement — whether country-based or company-based — remains unresolved.

What is the EU seeking from Anthropic?

The European Union is seeking access to Mythos 5 specifically to study the model's implications. The EU's interest reflects broader concern among G7 nations about being locked out of advanced AI tools used for cybersecurity defense.

An agreement providing broader access would allow G7 countries to use the models to build stronger cybersecurity defenses against rivals such as China, according to reporting by IndexBox.

Which AI executives are attending the G7 summit?

AI executives from Anthropic, OpenAI, and Google are expected to attend a working lunch on Wednesday at the summit. The lunch agenda covers technology issues including regulation, AI infrastructure, and networks.

The presence of Altman, Amodei, and Hassabis at the G7 underscores how central AI governance has become to this year's summit agenda. Reuters previously reported the Wednesday lunch.

Anthropic's spokesperson did not immediately respond to requests for comment.

Timeline of key events

What sectors lost access to Mythos?

Before the Trump order, Anthropic had granted Mythos access to organizations across multiple critical infrastructure sectors. Those sectors included:

• Healthcare
• Communications
• Power
• Water

Those organizations used Mythos to scan their own systems for vulnerabilities. That access was cut when Anthropic disabled the model for all users last Friday.

The OpenAI deployment safety debate and the Anthropic restrictions both point to a broader pattern: governments are increasingly shaping how frontier AI models reach international users.

With AI infrastructure spending rising sharply in 2026, the stakes around who gets access to frontier models — and under what conditions — are growing for every G7 nation.

The next confirmed milestone is the Wednesday working lunch in Evian-les-Bains, where executives from Anthropic, OpenAI, and Google are scheduled to meet with G7 leaders on AI regulation and infrastructure.