New Delhi, June 17, 2026: G7 nations are considering a proposal that could allow select allied countries and organizations to access some of the most advanced artificial intelligence models developed by U.S. technology companies.
The discussions reportedly took place during the G7 summit, where leaders and officials explored ways to balance national security concerns with international cooperation in the rapidly evolving AI sector. The proposal is aimed at creating a framework through which trusted partners can gain access to advanced AI capabilities while maintaining strict security safeguards.
The development comes shortly after tighter restrictions were introduced on access to cutting-edge AI systems. Anthropic recently limited foreign access to its latest AI models, Fable 5 and Mythos 5, following U.S. government directives focused on protecting sensitive technologies and addressing national security risks.
According to sources familiar with the matter, the proposed “trusted partners” framework could include both countries and selected organizations that meet specific security and compliance standards. The initiative is intended to ensure that key allies continue to benefit from advanced AI tools without compromising strategic interests.
Supporters of the proposal believe broader access to sophisticated AI models could significantly strengthen cybersecurity capabilities among allied nations. These technologies can help identify vulnerabilities, improve digital defenses, and protect critical infrastructure from increasingly complex cyber threats.
The U.S. administration has indicated that it remains engaged with partner nations on issues related to AI security and technological cooperation, while continuing efforts to address concerns surrounding the use and distribution of advanced AI systems.
Executives from major AI firms, including Anthropic, OpenAI, and Google, are expected to participate in discussions focused on AI regulation, infrastructure development, and the future of global technology governance.
Cybersecurity experts have highlighted the capabilities of Anthropic’s Mythos model, which is designed to identify weaknesses in software code and computer systems. The technology is considered valuable for organizations seeking to enhance their cyber defense mechanisms and strengthen network security.
Before the recent restrictions were introduced, Anthropic had provided access to Mythos to selected organizations across more than 15 countries. These organizations operated in sectors such as healthcare, communications, energy, and water services, using the model to detect and address security vulnerabilities within their systems.
As discussions continue, the proposed trusted-partner framework could play a key role in shaping future international cooperation on artificial intelligence, while ensuring that security considerations remain a central part of AI policy and governance.
Author
About the Author
