I graduated high school back in 1998. In those days, university academics were waking up - albeit slowly - to the potential the Internet had for research. But my high school English teacher hadn't turned the corner. Anything discovered on the Internet didn't count as a valid source because "anyone can say anything on the Internet." It simply didn't matter to her in those days that legitimate, university hosted info existed online.
We can look back on such skepticism and smile now. Most research is done online, and the sources they link to are online as well. Print media dwindles.
At the same time, we still need a healthy dose of skepticism. Examining sources is critical. "The most essential gift for a good writer," Hemingway wrote, "is a built-in, shockproof, shit detector."
AI doesn't have one. Even the best of models go scraping the web indiscriminately, finding whatever sources it can. My favored Claude 4 Opus does it. Elon's robot does it. Thanks, but no thanks, Grok: I do not want you to scan Quora for my test question regarding the nature of Christ, any more than I'd like you to accept the Gospel Coalition as a valid source.
Turns out, anyone can in fact say anything on the Internet.
Perhaps in the future, LLMs will set higher standards for the information they incorporate into answers, especially as it pertains to academic questions. For the time being, however, keep your detector set to high.