All micro-internships
Micro-internship brief ≈4h in one sittingno-code

Recreate a photo you take with AI

Take an original photo of any scene or object — it does not need to show your face — then recreate it with an AI image tool through at least 5 prompt iterations. The gap between your photo and the AI’s best attempt reveals what these tools can and can’t do.

AI LiteracyPrompt Engineering

Posted by The AI Internships

The work brief

  1. 01Take an original photo with your phone. Pick one with specific details: objects, lighting, textures, and details specific to your own place or culture.
  2. 02Describe it to an AI image tool and generate. Compare honestly against the original.
  3. 03Iterate at least 5 times, changing your prompt deliberately each round — composition words, lighting terms, style anchors.
  4. 04Stop when you plateau. Analyse the remaining gap: what could you never get the AI to do?

What you’ll produce

4 deliverables

Submission standard

Submit your original photo, the AI’s best attempt, all prompts in order, and the gap analysis. The original must be clearly a real phone photo — that’s the anchor of the whole exercise. Upload a second photo of the same scene from a different angle so we can tell your original is a real photograph.

  • Your original photo (taken by you)

    Must be an un-edited phone photo of a real scene with a specific personal/local detail; AI-looking 'originals' fail.

    Image uploadRequired
  • The AI’s closest attempt

    Image uploadRequired
  • All your prompts in order, with what you changed each round and why

    Written responseRequired
  • The gap: what the AI got right, what it never could, and your theory why

    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.

  • Second photo of the same scene from a different angle

    A second real photo of the SAME scene from a different angle or moment — this proves it's a real 3-D scene, not an AI image.

    Image uploadRequired

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.

Deliberate iteration

29%

5+ prompts that show strategy, not random retries.

Result quality

14%

The best attempt is impressively close given tool limits.

Gap analysis

29%

The analysis names specific, observed limitations — not generic "AI isn’t perfect".

The original is a real photo

29%

Auto-fail if the 'original' looks AI-generated, or the second angle doesn't depict the same real scene (invented anchor).

How we grade your AI usage

30% of your score

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.