Y’all. This is wild. Yesterday I had a moment — this thing had happened and as I processed it I decided to post about it on LinkedIn. I figured the usual suspects would like it, maybe someone would comment. I wondered if I should post about it at all, figured I could always delete later if I wanted to.
I did not anticipate this kind of activity, with hundreds of reactions, 50+ comments, shares in the double digits, and 18 new connections.
Full post:
This is unsettling in the best way—the kind of post that makes me sit up. The phantom citation is almost poetic: a paper that sounds like you, built from real fragments, but hollow at the center. That's the part that gets me. Hallucinations don't look like errors; they look like plausibility.
ReplyDeleteWhat I keep landing on is that the safeguard isn't the tool, it's the discipline behind it. I use AI constantly—I'll ask Claude to argue against my own thinking or pressure-test a claim—but the moment it becomes a shortcut around reading the actual source, the whole thing collapses. The work is the point.
Your closing line stays with me: if we're not finding real sources and writing our own manuscripts, it's worth asking why. I'd add—worth asking before the temptation arrives, not after. Thank you for naming this so clearly.
This is exactly what I’m talking about! I only have one request: please keep posting, and post even more!
ReplyDeleteSocial media needs voices like yours. We need thoughtful perspectives that spark meaningful conversations about the dilemmas and challenges we are facing today.
Thank you for raising this so publicly. When I came across this post last week, I was struck by how visibly academic integrity is eroding and how little scrutiny AI-generated output is receiving before it enters the scholarly record. The volume of engagement on your post made clear that this is not a peripheral concern, it is something the academic community recognizes as urgent and worth addressing now.
ReplyDelete