Anthropic revealed on November 6 (local time) the results of a "sabotage" experiment in which a covert disruptor was planted inside an AI team. Sabotage refers to pretending to cooperate while secretly undermining work behind the scenes.
