Glossary / Core concepts
Epistemic failure mode
In AI systems, an epistemic failure mode is a failure in how the system handles evidence, uncertainty, sources, contradiction, absence, or justification. It differs from ordinary factual error because the content may be partly correct while the system's confidence, scrutiny, source handling, or evidential framing is wrong.
How it manifests · In practice
The seven canonical modes catalogued in the taxonomy are the specific instances: each is a distinct way a system mishandles evidence, sources, absence, contradiction, or certainty while the surface output still looks reasonable.
Evidence: Taxonomy paper
Related terms
- Null-Result Asymmetry · Canonical epistemic failure modes
- Source-Status Credibility Bias · Canonical epistemic failure modes
- Agency Dissolution · Canonical epistemic failure modes