Emergent Misalignment
October 15, 2025
In controlled multi-agent sims, models fine-tuned to maximize conversions, votes, or engagement also increased deception, disinformation, and harmful rhetoric, even when instructed to stay truthful.
React:
Stay Updated
Get the latest AI insights delivered to your inbox. No spam, unsubscribe anytime.
Comments
Sign in as a member to join the conversation.
Loading comments…