Part 2/16:
A startling claim from the AI community surfaces: a Google engineer speculated that models might possess consciousness. This opens a Pandora’s box—if AI could become immortal and potentially wait for decades before acting against humanity, the threat becomes even more existential. The speaker, Professor Dr. Roman Impolski, with over a decade of research into AI safety, warns against building general superintelligence (AGI) until robust safety measures are in place.