
Anthropic revealed breakthrough research using “persona vectors” to monitor and control artificial intelligence personality traits, introducing a counterintuitive “vaccination” method that injects harmful behaviors during training to prevent dangerous personality shifts in deployed models.