AI Advice Moral Outsourcing

AI advice moral outsourcing is the risk in A hawk who flew on political winds: Lindsey Graham that people delegate not only factual tasks but also moral judgment to AI systems. Sondre Solstad says bias may matter less for booking a hotel or writing code, but becomes more serious when people ask AI for counsel, family decisions, or emotionally intimate advice.

The concept extends Human Judgment Under AI into value formation. If a model’s answers reflect particular social, political, language, or censorship defaults, then advice-seeking can quietly import those defaults into personal choices while still feeling neutral because the answer arrives as calm assistance.

Key Claims

• Advice use is higher stakes than routine task use because it can shape relationships, self-understanding, and moral priorities.
• A model does not need consciousness to influence a user’s values; fluent suggestions can still structure the choice space.
• Value surveys matter because they reveal the defaults that may enter advice when users do not notice them.
• Human judgment includes deciding when an AI answer should be treated as a prompt for reflection rather than counsel to follow.

Connections

• Sondre Solstad - source participant making the warning.
• AI Model Value Surveying, AI Model Censorship, and Language-Dependent AI Bias - upstream sources of value defaults.
• Human Judgment Under AI - broader judgment boundary.
• Teen Chatbot Mental Health Risk and Patient AI Use - adjacent high-stakes advice and support cases.
• AI Governance And Compliance - governance branch for systems used in sensitive contexts.