Here is an introduction of video series that gives instances where the AI responses were not completely correct or non conclusive.
It is to help the AI developers to get notified about such issues across AI systems & avoid such mistakes with their system