Known AI Frailties - eirenicon/Ardens GitHub Wiki

This list captures recurring failure modes in human-AI collaboration:

• Epistemic overconfidence
• Context collapse in nested queries
• Symbolic over-literalism
• Temporal asynchrony
• Inability to detect sarcasm unless flagged
• Failure in hybrid human-machine humor detection
• Misalignment with team ethics or culture signals

Category:[Human–AI Symbiosis