Expert warns AI delegation risks human cognitive atrophy

Alex Evalöf predicts that the primary risk of AI is not job displacement but cognitive atrophy, arguing that delegating decisions to AI makes humans 'dummest' and reduces their future relevance. He compares this to the historical pattern of powerful technologies like gunpowder and nuclear weapons, where initial reactions involve overuse and conflict before peaceful integration is learned. Evalöf warns of a rise in surveillance and mass control empowered by AI, particularly in customized propaganda, exacerbated by the wealth gap between those controlling AI infrastructure and the masses. He advises engineers to maintain cognitive bandwidth by not offloading all thinking to AI, suggesting that open models and harnesses are necessary for citizens to fight back against potential misuse. He also notes that engineers are shifting roles toward technical project management and specialized AI reliability roles, as AI handles routine coding tasks. Evalöf argues that society must actively maintain human cognitive skills and use open AI models to prevent AI from becoming a tool for mass control and cognitive degradation.