# CrewAI YT Video: “How to Detect AI Hallucinations with CrewAI and Opik” **URL:** \ **Category:** Announcements\ **Created:** [May 6, 2025, 1:12am UTC](https://community.crewai.com/t/crewai-yt-video-how-to-detect-ai-hallucinations-with-crewai-and-opik/5658 "2025-05-06T01:12:26Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![maxmoura](https://sea1.discourse-cdn.com/flex025/user_avatar/community.crewai.com/maxmoura/32/4206_2.png) [@maxmoura](https://community.crewai.com/u/maxmoura)\ **Post date:** [May 6, 2025, 1:12am UTC](https://community.crewai.com/t/crewai-yt-video-how-to-detect-ai-hallucinations-with-crewai-and-opik/5658/1 "2025-05-06T01:12:26Z")
**Title:** How to Detect AI Hallucinations with CrewAI and Opik **URL:** [https://www.youtube.com/watch?v=IJ\_Bb0Vmve8](https://www.youtube.com/watch?v=IJ_Bb0Vmve8) **Channel:** Lorenze Jay (@Lorenze_Jay) **Summary:** Check this out if you want to evaluate your AI agents in CrewAI and catch those annoying hallucinations: the trick is to use CrewAI’s event listeners along with Comet’s Opik library. Basically, you set up a listener that kicks in whenever a task finishes, grabbing all the details – like the description, output, context, and even the tool calls. Then, you feed that data into Opik, which gives you a quality score and suggestions for improvement. And the coolest part? It’ll tell you if any hallucinations popped up by comparing the agent’s response against the context it was given. It’s super flexible and open-source, so you can tweak it for whatever task or workflow you’re running. ---
**Author:** ![zinyando](https://sea1.discourse-cdn.com/flex025/user_avatar/community.crewai.com/zinyando/32/4435_2.png) [@zinyando](https://community.crewai.com/u/zinyando)\ **Post date:** [May 6, 2025, 6:46pm UTC](https://community.crewai.com/t/crewai-yt-video-how-to-detect-ai-hallucinations-with-crewai-and-opik/5658/2 "2025-05-06T18:46:16Z")
Great content as usual from Lorenze 👏🏾