EU Commissioner warns AI oversight fails on self-improving models

European Commissioner for Human Rights stated at the University of Oxford AI Summit that current regional regulations are insufficient against recursive self-improving AI systems. The Commissioner warned that rapid technological developments threaten fundamental oversight principles, questioning whether transparency remains possible with such models. She highlighted a critical vocabulary gap between industry safety frameworks and formal human rights law, urging technical standardization bodies to embed public values directly. The address emphasized that European instruments like the AI Act cannot regulate global actors or defense applications operating outside jurisdictional boundaries. Furthermore, she called for urgent capacity building within oversight institutions and multidisciplinary dialogue involving civil society. The Commissioner also identified environmental impacts and energy resource competition as unaddressed gaps in current AI governance frameworks requiring immediate international attention.