French President Emmanuel Macron said he expects progress in the coming weeks on efforts to expand international access to Anthropic's advanced Mythos artificial intelligence model, placing AI cooperation at the center of discussions among world leaders at the G7 summit in France.

 

The comments came as leaders from the Group of Seven nations concluded talks in Évian-les-Bains, where advanced artificial intelligence emerged as one of the most significant topics alongside trade, security, and economic growth. 

 

Governments are increasingly treating frontier AI systems as strategic assets, creating tensions between national security concerns and the desire for broader international collaboration.

 

At the center of the debate is Anthropic's Mythos platform, one of the most powerful AI systems developed for advanced cybersecurity analysis and software vulnerability detection. Access to the technology became a major international issue after U.S. authorities moved to restrict foreign access to Anthropic's most advanced models, citing national security concerns. The decision led Anthropic to suspend access to both Mythos and Fable models while it worked to comply with government directives.

 

Macron indicated that discussions among G7 leaders had focused on finding a pathway that would allow trusted allies and partner nations to access cutting-edge AI technologies without compromising security requirements. 

 

According to diplomatic sources familiar with the talks, officials discussed a potential "trusted partners" framework that could provide selected countries and organizations with access to advanced American AI systems.

 

The proposed arrangement reflects growing concern among U.S. allies that restrictions on frontier AI models could create barriers to technological cooperation. European leaders have argued that access to advanced AI tools is becoming increasingly important for cybersecurity, scientific research, economic competitiveness, and national digital infrastructure.

 

Macron warned that limiting access to leading AI technologies could undermine confidence in American technology providers among international partners. His remarks highlighted a broader debate emerging across Europe about technological dependence and the future balance of power in the global AI industry.

 

The issue has gained urgency because Mythos is not viewed as a typical consumer AI product. Anthropic originally introduced the system as a specialized cybersecurity model capable of identifying software vulnerabilities and strengthening digital defenses. Before its wider release, access was limited to selected organizations participating in security-focused testing programs.

 

Industry leaders also participated in G7 discussions. Executives from Anthropic, OpenAI, and Google DeepMind met with government officials to discuss AI safety, governance, infrastructure, and international cooperation. Several executives emphasized the need for democratic nations to coordinate their approach to AI development rather than adopting fragmented national strategies.

 

European Commission President Ursula von der Leyen echoed similar concerns, saying it is in the mutual interest of both the United States and Europe to ensure access to advanced AI technologies while maintaining strong safeguards. She also stressed the importance of close transatlantic cooperation as governments develop frameworks for managing increasingly powerful AI systems.

 

The discussions come during a period of rapid change in the AI industry. Governments are increasingly treating frontier AI models as technologies with strategic importance similar to advanced semiconductors, cybersecurity tools, and critical infrastructure. As capabilities improve, policymakers are facing growing pressure to balance innovation, security, and international cooperation.

 

No final agreement was announced at the summit, but diplomats indicated that conversations around a trusted partners program are continuing. Macron's comments suggest that negotiations remain active and that additional announcements regarding international access to advanced AI systems could emerge in the coming weeks.