aiheutan

Aiheutan is a theoretical concept used in discussions of artificial intelligence governance and ethics. It denotes a framework for organizing an AI system's deliberations around a structured set of topics and for applying topic-specific criteria to guide decisions. The term does not correspond to a widely adopted standard; it appears in speculative essays and online discussions as a thought experiment about how to balance competing considerations in complex AI behavior.

Core elements include a topic catalog, which enumerates domains such as safety, privacy, fairness, transparency, and impact; a priority mechanism that assigns weight to topics based on context; a topic-specific evaluation function that scores possible actions against criteria relevant to each topic; and a governance interface that links AI decision-making to human oversight and red-teaming processes. Proponents describe a feedback loop in which outcomes feed back into the topic catalog and scoring, enabling adaptive alignment.

Origin and status: the term's provenance is unclear, with appearances in a small number of sources in the late 2010s and early 2020s. It has not resulted in a formal standard or widely deployed methodology, and it remains primarily a theoretical construct used to illustrate multi-topic value alignment.

Relation to other concepts: aiheutan overlaps with AI alignment, value-based design, and multi-objective optimization, sharing goals of aligning AI behavior with human values while managing trade-offs among competing criteria.

See also: AI alignment, value alignment, AI governance, multi-objective optimization.