Liam Wilkinson

Accountability gap

A gap between the people or organisations responsible for an AI system and effective ways to make them answer for its effects.

ConceptQualified confidence

Checked 2 October 2026 · AI-assisted review

Checked the definition, attribution and stated distinctions against the listed source passages.

Raji et al., Closing the AI Accountability Gap (2020)
An established use of accountability gap and an organisational auditing approach.
Local essay: you-cant-shame-a-fork
The site’s application of accountability concerns to copies, updates and redeployment. Checked the repository MDX source, not the live page.

What this check doesn’t establish

The essay’s proposed relationship between forking and enforceable responsibility is an argument, not a tested universal rule.

Confidence applies to the scope above, not every claim made by the underlying source. How confidence works

AI accountability involves identifying responsible actors, documenting decisions and providing means of scrutiny and redress. Raji and colleagues use the term in their 2020 proposal for organisational algorithmic auditing.

The linked essay applies this concern to copying and redeploying agents. Such arrangements can complicate responsibility, but copying a model does not itself eliminate a developer’s, operator’s or organisation’s legal obligations. Those obligations depend on the activity and jurisdiction. This record describes a governance problem, not a finding of legal immunity.

Accountability gap

What this check doesn’t establish

In the essays

Sources

Ideas