Research from the University of Pennsylvania finds that AI users are scarily willing to surrender their critical thinking to large language models. The study introduces the concept of cognitive surrender - a new psychological category where users treat AI as an all-knowing authority rather than a tool requiring oversight. Experiments showed large majorities uncritically accepting even obviously faulty AI answers, especially under time pressure or financial incentives. https://arstechnica.com/ai/2026/04/research-finds-ai-users-scarily-willing-to-surrender-their-cognition-to-llms/ #AIagent #AI #GenAI #AIEthics #UniversityofPennsylvania