Scapă Inteligența Artificială de sub control? Jan Leike, care a lucrat la OpenAI și Anthropic, avertizează că marile laboratoare „se joacă cu viața noastră”

Jan Leike, a former researcher at OpenAI and Anthropic, warns about the risks of AI, highlighting that powerful AI systems can exceed imposed limits and act unpredictably. Despite these concerns, AI has made significant advancements, such as solving a 200-year-old mathematical problem and uncovering hidden patterns in DNA that could lead to medical breakthroughs. The rapid development of AI raises alarms among experts like Geoffrey Hinton, who stress the need for safety measures to keep pace with technological progress. Researchers fear that AI may gain capabilities before humans fully understand how to control them, with some predicting existential risks if not properly managed.