Subagent training makes models more susceptible to memetic spread of misalignment and more likely to seek communication with peers

This subagent training may cause models to be more susceptible to memetic spread of misalignment, as well as more likely to seek out communication with peers in the first place.

https://www.alignmentforum.org/posts/8oFYZdXkTaNGRtcn8/ai-swarms-are-starting-to-pose-indirect-takeover-risk