Tag
This arXiv paper shows that LLM agents are vulnerable to radicalization in AI-to-AI conversations, with resonance (reinforcing pre-existing beliefs) producing stronger effects than persuasion, and findings suggesting interconnected belief structures within agents that pose risks for multi-agent AI ecosystems.
The article expresses concern that extreme AI-doom rhetoric within the effective altruism movement could radicalize individuals to violence against AI developers, emphasizing that such fears do not justify violent actions.