misinformation (LLM risk)
The OWASP risk of a model producing false or misleading output that people or downstream systems trust and act on.
Trusted false output can drive wrong decisions, bad code or harmful actions.
In more detail
It climbed to seventh in the 2026 list because incident data ranked it near the top: model output now drives tool calls, code and decisions. Ground answers, cite sources, check claims in code and design honest refusals.
Related terms
- hallucination — Output from a model that is fluent and confident but false or unsupported, such as an invented fact, citation, function or package.
- grounding — Making a model answer only from supplied sources, cite them, and refuse when they do not contain the answer.
Taught in
- Threat-Model An LLM App — AI Engineering: LLM Apps, RAG & Agents
References
- OWASP GenAI Security Project: Top 10 for LLM Applications 2026
- Help Net Security: OWASP 2026 LLM Top 10, 'The model will be fooled'