French President Emmanuel Macron used the G7 summit to argue that allied nations should get access to Anthropic’s Mythos AI, a move that pits national security interests against the push for open international collaboration. The appeal, which came during closed-door sessions, underscores the growing friction over how advanced artificial intelligence systems are shared among close allies.
What Macron proposed
Macron’s pitch centered on granting allied governments—not just the companies that build them—access to Anthropic’s Mythos AI. Mythos is considered one of the most powerful large language models in private hands, and its capabilities have drawn attention from defense and intelligence communities. The French leader argued that restricting access to only a handful of nations or corporations could destabilize alliances and create asymmetries in AI power.
Under the plan, allied nations would receive some level of visibility or use rights, though the exact terms weren’t detailed in public remarks. Macron framed the proposal as a way to keep strategic partners on equal footing while avoiding a fragmented landscape where each country races to build its own sovereign AI.
Security vs. collaboration
The tension between protecting national security and fostering international R&D has been a recurring theme in AI governance. Anthropic, the U.S.-based company behind Mythos, has its own security protocols and has been cautious about who gets access to its most advanced models. Macron’s push challenges that approach by suggesting that allied governments should have a seat at the table, not just the firms and their home countries.
Some G7 delegations expressed concerns that broadening access could lead to leaks or misuse, even among allies. Others saw the proposal as a necessary step to prevent a few countries from monopolizing AI breakthroughs. The debate highlights how quickly the technology is outpacing the diplomatic frameworks meant to manage it.
Calls for clearer rules
Macron’s advocacy also gave fresh urgency to the need for clear regulations governing AI development and access. Currently, no international treaty or binding agreement spells out how allied nations should share advanced AI systems. The French president wants the G7 to start drafting principles that balance openness with safeguards.
No formal decision came out of the summit. Instead, leaders agreed to task a working group with studying the idea and reporting back before the next G7 meeting. That leaves the question of how—and whether—Anthropic’s Mythos AI will be shared among allies unresolved for now.
