Meta’s new Incognito Chat with Meta AI is the first mass-market deployment of a chatbot the provider cannot read. We examine the architecture, the moderation trade-off, the liability consequence, and what it means for the Going Dark debate

The health agent rush may be the most consequential AI development of 2026. Yet, it has not received serious academic or regulatory attention. Until now.

Hundreds of millions of people confide their most intimate secrets to AI chatbots every day. The interface invites intimacy; the fine print reserves broad rights most users will never read. This first-of-its-kind study maps what really happens to your words across ChatGPT, Gemini, Claude, Grok, and DeepSeek. Read more

In his article published by the IAPP, Professor Christakis examines the “China data question” in greater depth, focusing on direct data collection, open-source deployment, and the limits of extraterritorial enforcement under EU law.

One year after “AI’s Sputnik moment,” the global regulatory response to DeepSeek has produced a starkly bifurcated outcome: constrained in the West, yet surging 960% worldwide. This comprehensive study documents the regulatory storm, unpacks the legal concerns, and confronts an uncomfortable truth about the limits of extraterritorial enforcement.

This study seeks to provide an overview of current progress in the designation of national competent authorities, and to analyse how national governance frameworks are being developed. The study shows that one week prior to the deadline, the majority of member states were behind schedule. A few draft governance frameworks, some of them unofficial, were announced. These present a centralised architecture for notifying authorities, but a fragmented one for market surveillance, with most authorities already identified.