PromptRiskDBThreat intelligence atlas
AI Risk

Knowledge conflicts in retrieval-augmented LLMs

"AI models can be particularly sensitive to coherent external evidence, even when they come into conflict with the models’ prior knowledge. This may lead to models producing false outputs given false information during the retrieval- augmentation process, despite only a relatively small amount of false informa- tion input that is inconsistent with the model’s prior knowledge trained on much larger amounts of data...

View related techniquesRead profile
AI Risk7. AI System Safety, Failures, & Limitations7.3 > Lack of capability or robustness2 - Post-deployment

Record summary

A quick snapshot of what this page covers.

Techniques0Attack methods connected to this risk.
Mitigations0Defenses that may help with related attacks.
Domain7. AI System Safety, Failures, & LimitationsThe broad risk area this belongs to.

Risk profile

How this risk is described and categorized.

"AI models can be particularly sensitive to coherent external evidence, even when they come into conflict with the models’ prior knowledge. This may lead to models producing false outputs given false information during the retrieval- augmentation process, despite only a relatively small amount of false informa- tion input that is inconsistent with the model’s prior knowledge trained on much larger amounts of data [220]."

Domain7. AI System Safety, Failures, & Limitations
Subdomain7.3 > Lack of capability or robustness
Entity2 - AI
Intent2 - Unintentional
Timing2 - Post-deployment
CategoryAttacks on GPAIs/GPAI Failure Modes
SubcategoryKnowledge conflicts in retrieval-augmented LLMs

Related techniques

Attack methods connected to this risk.

No linked attack methods. No AI attack method is connected to this risk in the current data.

Suggested mitigations

Defenses that may help with related attacks.

No propagated mitigations. No defense is available through the connected attack methods.

Source

Research source for this risk, when available.