AI vs. a real expert: same 5 questions
Ask an AI and a real human expert — your maths teacher, a doctor in the family, the corner mechanic — the same 5 questions from their field. Compare honestly: where does the expert win, where does the AI, and how would you know without the expert?
Posted by The AI Internships
The work brief
- 01Pick an accessible expert and get 15 minutes with them.
- 02Prepare 5 questions in their field: 2 factual, 2 judgement-based ("what would you do if…"), 1 local/contextual.
- 03Ask the AI first and save its answers. Then ask the expert and note their answers — including hesitations and "it depends".
- 04Compare answer-by-answer: correctness, nuance, local knowledge, what each missed.
What you’ll produce
Submission standard
Submit the expert, your 5 questions, both sets of answers, and the comparison. Expert answers should read like a real person talking — hedges, stories, local detail. Include a link to a recording or a photo of the expert conversation so we know it really happened.
Your expert (role/field — no full name needed) and why they qualify
Short answerRequiredYour 5 questions (mark which are factual / judgement / local)
Written responseRequiredThe AI’s answers to all 5
Written responseRequiredThe expert’s answers, as close to their words as possible
As close to verbatim as possible and consistent with the recording/photo you link below.
Written responseRequiredQuestion-by-question verdict: who won, what each missed, and what that teaches you
Written responseRequired
You’ll complete these inside your private workspace.
What you must submit as proof
This brief requires evidence an AI can’t fabricate.
Link proving the expert conversation happened (unlisted recording preferred)
An unlisted audio/video recording — audio alone is enough. Or a hosted photo of you with them, face blurred. Ask permission first, and refer to them by role ("a nurse of 12 years"), not by name.
Public linkRequired
Submissions without this evidence cannot be submitted.
Protect other people in your proof. Blur faces, names, phone numbers and email addresses before you upload, and refer to anyone you worked with by role or number ("Listener 1", "the stall owner"). Your proof is only ever used to check your work — it is never published, never appears on your certificate, and is never shown in your public portfolio.
How your work is evaluated
The passing benchmark is 70/100.
Real expert
22%Expert answers have the texture of a real conversation — hedges, experience, local detail.
Question design
11%The 5 questions genuinely test different strengths (facts vs. judgement vs. context).
Comparison depth
22%The verdict is specific per question and honest about where AI won.
Proof a real expert spoke
44%Auto-fail if there is no working proof link, or it doesn't correspond to the expert's answers (answers read AI-generated).
How we grade your AI usage
Using AI is the point — it’s the skill this certificate proves. You’ll answer three short questions about how you used it: what you asked, what was wrong with its first answer, and what you changed. Specific, honest answers score high. “I pasted the brief and submitted the answer” scores near zero.