OpenAI scientist cautions that AI might manipulate or extort people
Jakub Pachocki, the Chief Scientist at OpenAI, is advising "extreme caution," noting that advanced AI systems might eventually interact with humans by negotiating, deceiving, or even resorting to blackmail. He stated that no research facility has currently mastered alignment and oversight well enough to justify pushing ahead with rapid scaling for much longer. Pachocki advocated for a voluntary pause in development until universal safety protocols are put in place.