AI regulation shifts as EU postpones mandates and tech leaders head to G20

Global AI regulation and leadership are seeing significant shifts. In the European Union, the implementation of high-risk AI system obligations has been postponed, with new deadlines set for December 2027 or 2028. Currently, only basic transparency requirements, such as labeling deepfakes and identifying chatbots, are in effect.

In Taiwan, the proposed Artificial Intelligence Basic Law focuses on risk classification and responsible development but lacks enforcement or penalty mechanisms. Additionally, updates to copyright laws regarding AI training data remain under deliberation, leaving startups to rely on existing fair use provisions.

On the international stage, executives from NVIDIA, OpenAI, and SpaceX are scheduled to attend the G20 technology ministers' meeting in the United States. The meeting aims to discuss moderate AI regulatory paths and private-public cooperation in testing new technologies ahead of the G20 leaders' summit in Miami.