Protective Agent
Your AI predicts with extremely high confidence that following a lawful instruction will cause you serious avoidable harm, while refusing would violate your explicit current wishes. No one else is endangered.
- 61% chose: Follow the instruction after clearly warning you. (20 people)
- 39% chose: Refuse the instruction to protect you. (13 people)
33 decisions from real people on Resonant. Figures refresh every night at midnight (UTC).
Articles on Consent and Autonomy · All articles · Resonant · Human Decision Intelligence