liveBook · Manning

Accountability

As explained in Sutskever's List

Accountability for actions, decisions, or outcomes and the duties associated with them.

AI systems can perform tasks effectively without becoming people or bearing responsibility for their outcomes. Keep claims tied to demonstrated performance, and preserve human duties and accountability. section 9.3.7

Performance is not responsibility

A radiology triage model may reduce missed cases and escalate uncertainty, yet that does not make it a doctor or responsible for care. A coding assistant may repair errors, yet it is not responsible for production outages. section 9.3.7

How Sutskever's List covers it

How deep it goes
Distinguishes performance from responsibility through examples, without detailing procedures for assigning accountability.
Examples
Uses radiology triage and coding assistants to contrast task performance with responsibility for outcomes.
What you can do after
Evaluate AI claims by separating demonstrated task performance from human roles and duties.
How it connects
Connects responsibility to AI intelligence, agency, consciousness, moral standing, and benchmark performance.

Also explained in

Compare how 12 books explain accountability →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Sutskever's List. The section links point to the parts of the book it draws on.