Of course, it begins and ends with Anthropic’s Mythos class of frontier AI models – the bone of all contentions in all things AI right now. The US government’s unilateral decision to restrict the export of Claude Fable 5 and Mythos 5 models for the rest of the world was a topic India commented on at the G7 2026 summit in France.
In PM Modi’s official statement at the G7 summit, he pointed out to world leaders – including US President Donald Trump sitting right next to him – how “no country can be completely secure in cyberspace until all countries are secure.” In this PM Modi was correct, if you remember the old adage how the chain is only as strong as its weakest link.
Continuing in his official statement, PM Modi further said how, “India has always viewed cyberspace as a Global Public Good. Consequently, access to these critical AI technologies must also be widespread and inclusive.”
PM Modi made India’s stance on who should (and shouldn’t) get access to leading edge frontier AI models like Anthropic’s Mythos class of models. “All democratic nations should have access to such AI models so that they can protect their critical information infrastructure and counter growing cyber threats.”
India believes access to frontier AI models should be “broad and inclusive.” For India, it isn’t so much a question of commercial opportunities but one of safeguarding fundamental building blocks of the digital world against cyber threats.
India placed four specific proposals before the G7 summit members, prime among which is the need to develop “safe-by-design systems.”
Also read: Fable 5 shutdown reveals a dangerous gap between AI regulation and AI reality
Safety shouldn’t be an afterthought in the development of any technology meant for mass deployment, PM Modi argued India’s position at the G7 summit. It moves the debate beyond post-release moderation. What would “safe by design” AI systems look like? Lots of model evaluations, access controls, secure development, misuse testing and protections against dangerous outputs, to name just a few starting points.
PM Modi called for G7 countries to develop common AI standards, shared testing framework, and regulatory sandboxes in which models and applications could be tested under supervision.
Speaking broadly about AI’s efficacy, PM Modi reminded world leaders and technical visionaries from the likes of Anthropic, OpenAI and Google DeepMind (who were in attendance) on how AI’s power shouldn’t be measured in machine capabilities but the technology’s ability to empower ordinary people.
Indian PM Modi also highlighted child safety as the moral case for global standards in AI, warning how the same AI that can educate and empower children and enhance their creativity can also expose them to misinformation, deepfakes and exploitation – in the absence of AI standards, governance and guardrails.
Also read: Anthropic removes access to Claude Fable 5 and Mythos 5, here is why