EU AI Act continues moving forward

On May 11, 2023, key committees in the EU Parliament approved compromise amendments to the draft AI Act, clearing the way for a plenary vote in the Parliament in mid-June and, should it be approved, finalization of the Act’s provisions. The EU Parliament has stated that the amendments were designed to “ensure that AI systems are overseen by people, are safe, transparent, traceable, non-discriminatory, and environmentally friendly.” In particular, the compromise amendments include new provisions limiting AI systems’ application to biometric identification and categorization, scraping of biometric data from social media or CCTV footage, emotion recognition and predictive policing. The amendments also expand the definition of high-risk areas to include “harm to people’s health, safety, fundamental rights or the environment” and “systems to influence voters in political campaigns.” The amendments also include stricter transparency measures that would likely apply to general-purpose AI systems (e.g., ChatGPT, Bard, Claude), including obligations to disclose that content was created by AI, to design models to prevent “generating illegal content” and to publish “summaries of copyrighted data used for training.” Once finalized and adopted, the Act will become applicable 24 months after entry into force (i.e., mid-2025 given current timelines).