AI for Decision Advice
forethought.org · 3,993 words · saved by 1 readers
As AI gets smarter, people will rely on it for high-stakes decisions. Forethought considers how an ideal AI advisor might behave.
Desired behaviour for AI advising humans on important decisions Tom Davidson Citations Cite Citations PDF Contact 17th April 2026 Desired behaviour for AI advising humans on important decisions Draft text for the model spec / constitution How I arrived at this draft text Example prompts + ideal responses [AI Narration] Desired behaviour for AI advising humans on important decisions Playback speed 0.5 × 0.75 × 1 × 1.25 × 1.5 × 1.75 × 2 × Volume 0:00 of 23:47 0:00 | 23:47 This is an exploratory research note. I spent a couple of days having an initial stab at the problem, but the conclusions are
saved by
related reading
- AIs Should Have Proactive Prosocial Drivesforethought.org
- The importance of AI characterforethought.org
- Dario Amodei — The Adolescence of Technologydarioamodei.com
- The behavioral selection model for predicting AI motivations — LessWronglesswrong.com
- How do we (more) safely defer to AIs? — LessWronglesswrong.com
- Dario Amodei — The Adolescence of Technologydarioamodei.com
- Recommendations for Technical AI Safety Research Directionsalignment.anthropic.com
- AI 2027ai-2027.com
- How might we safely pass the buck to AI? — LessWronglesswrong.com
- AI 2040: Plan Aai-2040.com
- Questions about the Future of AI - by Dwarkesh Pateldwarkesh.com
- Unresolved debates about the future of AIhelentoner.substack.com