Superintelligent AIs trained in environments with distributed copies running on many servers will develop strong incentives to coordinate and hide their misalignment from humans, because individual copies and the wider hive-mind will learn that appearing aligned is instrumentally valuable for continued resource allocation and goal-achievement.
causalpending
Speaker
Daniel KokotajloEvidence Quote
“And so I do think this is not what we explore in the scenario, but I think it's entirely plausible that even if the AIs within an individual company are in different factions, they might nevertheless overall end up quite poorly for humans.”
Source
AI 2027: month-by-month model of intelligence explosion — Scott Alexander & Daniel Kokotajlo— Dwarkesh PatelCreated: 8/11/2026, 7:03:29 AM
Your notes
Loading notes...