ConceptSelf-Evaluation Bias
Self-Evaluation Bias
The inherent tendency of an AI model, when evaluating its own output, to follow the assumptions and reasoning paths used to generate that output, thereby systematically overlooking errors of the same kind. It is one of the core problems in AI-assisted code review that requires independent verification mechanisms to address.
Timeline (last 90 days)
Oct 12
When a model that generated code reviews its own output, it tends to follow the same assumptions and reasoning paths that produced the code, and may systematically overlook the same kinds of errors, which is called self-evaluation bias.
Unverified50%