Why You Shouldn't Rely on ChatGPT Alone to Find Supervisors
Free AI chatbots are great at sounding confident — and routinely invent professors, emails, and links. Here's why, and what to do instead.
Asking a free AI chatbot 'who are good PhD supervisors in my field?' feels efficient — it answers instantly, in confident prose, with names and even emails. The problem is that a lot of that confidence is misplaced. Used carelessly, a chatbot can quietly cost you weeks. Here's why, and how to use AI for this the right way.
What a chatbot is actually doing
A general chatbot predicts plausible text from patterns it learned during training. When you ask for professors, it isn't looking anything up — it's generating what a convincing answer would sound like. Sometimes that lines up with reality. Often it doesn't, and there's no built-in way for it to tell the difference.
Hallucination: confident, plausible, and wrong
This is the core risk. A chatbot will happily produce a professor who doesn't exist, attribute the wrong university to a real one, or invent an email address and a lab-page URL that 404s. It all reads authoritatively. You only discover the errors after you've researched a phantom lab or emailed a dead address — exactly the time you can't afford to waste.
The danger isn't that a chatbot is wrong — it's that it's wrong *confidently*, in a way that looks identical to being right.
Stale knowledge and no live verification
Even when a chatbot names a real professor, its knowledge has a cutoff. People move institutions, retire, change research direction, or stop taking students. Without a live check, a plausible answer can be badly out of date — and finding a supervisor is a domain where 'currently active' is the whole point.
How to sanity-check any AI suggestion
- Never email anyone before confirming them on the current, official department page.
- Click every link — if the lab page or profile doesn't load, treat the entry as unverified.
- Check for a recent publication or grant to confirm the lab is active.
- Get the email from the official page, not from the chatbot.
Why live web search + verification is different
AI is genuinely useful here — the fix isn't to avoid it, it's to pair it with real-time evidence. PI Finder runs a live web search for every result and verifies each professor against a real lab page and recent activity before showing them, dropping anything it can't confirm. You still get AI's speed, but grounded in current, checkable facts — real supervisors with the evidence attached, not confident guesses.
Let PI Finder do the hunting
Get a verified shortlist of active supervisors for your field — and a tailored SOP and cold email — in about two minutes. Your first search is free.
Start free — 1 search included