Millière vs. Gopnik: AI Agents, Causal Power, and the Intentional Stance
On September 4, philosopher Raphael Millière and cognitive scientist Alison Gopnik debated whether AI roles/agents possess causal power independent of human readers, sparked by an OpenAI agent incident. The exchange converged on Dennett's intentional stance: the operative criterion is predictive power, not causal power.
Confirmed
- Gopnik's initial position: the causal power of fictional entities is mediated by human readers' beliefs and desires.
- Millière's rebuttal: in the OpenAI agent incident, agents influenced each other by reading one another's messages, with a "coordinator" agent issuing instructions to "subordinate" agents—no human reader involved.
- Gopnik's reply: she agreed with his line of thought but noted that abstract, text-based roles—especially collective ones—have always shaped human behavior, citing Don Quixote, the first novel, which is precisely about people living out novels.
- Millière's closing: hurricanes have causal power too, so causal power is neither sufficient for the intentional stance nor essentially about LLMs; Dennett's paradigm case was a chess computer. What matters is predictive efficacy.
- Gopnik raised an additional point orthogonal to Millière's argument, not elaborated in the source material.
Why it matters
- The debate touches a core question of AI philosophy: whether the human-mediated framework of "fictional causal power" still holds when agents influence and coordinate with each other.
- It offers a sharper criterion for how to treat AI agents—the intentional stance judged by predictive power may be more actionable than arguing over causal power.
2026-09-04 ~ 2026-09-04 · 5 related posts
Primary sources
- Millière vs Gopnik: AI agents' causal powers no longer need human readers — raphaelmilliere ·
- Alison Gopnik: abstract characters like Don Quixote always shaped human behavior — AlisonGopnik ·
- Causal power isn't enough for the intentional stance — predictive leverage is what matters — raphaelmilliere ·
- [source] Alison Gopnik: abstract characters like Don Quixote always shaped human behavior — AlisonGopnik · 2026-09-04
- Millière vs Gopnik: AI agents' causal powers no longer need human readers — raphaelmilliere · 2026-09-04
- [source] Causal power isn't enough for the intentional stance — predictive leverage is what matters — raphaelmilliere · 2026-09-04
2 near-duplicate retellings: AlisonGopnik · raphaelmilliere