And that while the web is full of citation-checking workflows by now, including using the very same LLM to help set things up so it's less of a chore.

You wouldn't trust LLM generated software without unit tests either.

(Maybe not even with just the unit tests, but you've got to start somewhere.)

Either way, pointing the finger at your tools is a cheap move. Don't blame the chainsaw for your own negligence!

Checking the citations is a loop away, zero excuses, they should be fined heavily for it as it just show reckless behavior, not "AI usage", using AI to produce an accurate document is not an issue in 2026.