The danger is that generative tools can amplify confirmation bias by enabling users to construct plausible explanations in domains where they lack expertise — reinforcing confidence without deepening understanding.
In medical contexts, this often manifests as self-medication. How many of us have already asked ChatGPT to diagnose the cause of a headache or interpret a harmless skin irregularity?