Viral ChatGPT photo prompts (and what makes them work)

The ones people keep using

• “What would I have looked like in the 80s?” — the 1980s restyle. Big hair, neon, synth backdrop. The exact prompt + a stronger variant: 80s ChatGPT prompt.
• Old-photo restoration — "restore and lightly colourise this old photo, keep it realistic, do not change faces". Great for family archives.
• The action-figure box — "make me as a boxed action figure with accessories". Peaked mid-2025, still circulating.
• Decade makeovers — the 70s, the 90s, "the most 1980s image ever created".
• Art-style transfers — Ghibli, oil painting, comic panel.

What the good prompts have in common

1. One clear photo as the anchor. Front-facing, well lit.
2. A specific target — a decade, a medium, a scene — not just "make it cool".
3. An identity guardrail — "keep my face and expression the same", "do not turn me into someone else".
4. A realism note — "keep it a photograph", "no text or watermark", "one person only".

Leaving out the guardrails is why results sometimes come back looking like a different person.

The limits

Every one of these needs a paid ChatGPT plan for image generation, and every one varies between runs. For the 80s specifically, a tuned one-tap tool skips the re-rolling — see ChatGPT vs a dedicated 80s filter.