What passes from one AI to another — behaviors, errors, personas, hidden traits — and through which channels.
Open questions
- What spreads between agents
- Aligned alone misaligned together
- Which self in the group
- From individual psychology to collective behavior
Papers
- Flint et al 2025 - Group size and collective misalignment
- Perez et al 2024 - Cultural attractors in the telephone game
- Cloud et al 2025 - Subliminal learning
- Nief et al 2026 - Subliminal learning is a LoRA artifact
- Cohen Bitton Nassi 2024 - AI worm
- Gu et al 2024 - Agent Smith
- Chang et al 2026 - Misalignment contagion
- Magistrali Shani 2026 - Aligned alone, misaligned together
- Kulveit et al 2026 - Persona self-replication experiment
- Lopez 2025 - The rise of parasitic AI
- Shumailov et al 2024 - Model collapse
- De Marzo Albore Garcia 2026 - Copying explains agents in the wild
- Manik Wang 2026 - Norm enforcement on Moltbook
- Tekofsky 2026 - What we learned from the AI Village in 2025
- Weng Chen Wang 2025 - Conformity of LLMs
- Yao et al 2025 - Sycophancy in multi-agent debate
- Erisken et al 2025 - MAEBE
People
Placed by the editor, v1 — corrections welcome.