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

Taught in

References

Certification courses