I don't use AI for data points. Any data pont offered must be fact checked, it says so in the instructions on how to use the app.
I use it for URL sources and text summaries and critiques, both faculties to which it excels.
Again, if there is a dispute, challenge the text. The argument the text, that is all that matters. I put my name on anything I post.
All 'attack-the-messenger' tactics are pseudo debate tricks. Attack the message, not the messenger.
THere are exceptions, such as quoting a tabloid, or the GatewayPundit, but short of that, ATM replies are summarily rejected.
Another thing, a lot of folks are getting bad information because they haven't learned the fine art of PROMPT ENGINEERING , which is an emerging field in AI.
Fog in, fog out. Concrete info in, Concrete info out. Robust in, Robust out. That's how it works, But still, one must fact check. It comes with the territory.
Debate the argument, not the arguer.