A new privacy treaty drafted by the Council of Europe extends the privacy framework beyond the EU, reaching 55 nations bound to the agreement.

The intro framing the draft notes governing AI chatbot memory, agent data access, and training-data handling, expanding privacy rules beyond GDPR to non-EU countries.

Formally, the framework is on track for adoption at the Council’s November plenary session.

The report on this development comes from TechTimes’ Scott McCain, dated September 4, 2026, citing the draft agenda and related links from the Council of Europe.

As a landmark update to AI privacy rules, the draft addresses how AI systems remember conversations, access data, and guard against training-data leakage, in a landscape predating modern internet-era norms.

The draft specifies rules on chatbot memory retention, autonomous agents’ access to corporate data, and handling potentially extractable training data to protect data privacy.

The guidelines were published on September 3, 2026, with reviews by the Bureau of Convention 108’s Consultative Committee planned for mid-September in Paris, aiming for formal adoption at the late November plenary.

This framework is part of the Council of Europe’s broader privacy and data protection agenda, aligning with global concerns about AI privacy and governance.