What ChatGPT Prompts Are Trending for Family Photo Edits Right Now

Straight answer: Right now the family photo edits doing the rounds on ChatGPT are Studio Ghibli-style portraits, 1990s disposable-camera recreations, "toy box" action figure edits, Renaissance oil painting portraits, and old-photo restorations that bring a missing grandparent back into the frame. I've run most of these on my own family photos this month, and the ones that get the most likes on Instagram are not the ones that hold up when you look closely at the faces.

If you want to go deeper on this: how to use chatgpt to write faster.

The five family photo prompts everyone's copying this month

I keep a running list of what's showing up in my feed and in client messages asking "how do I make my family photo look like that?" These five are the ones I've seen repeated word for word, which tells me they're being copied from the same few viral posts rather than written fresh.

1. The Studio Ghibli family portrait

This one is still going strong, over a year after it first blew up. The prompt people paste in is usually some version of "Turn this photo into a Studio Ghibli-style anime illustration, keep the same poses, expressions and clothing, soft warm lighting, hand-painted background." It works best on two or three people. I've written a full breakdown of exactly how I word this one for a Ghibli-style family portrait if you want the wording that gives you faces you recognise rather than generic anime characters.

2. The 1990s disposable camera photo

This is the newer one, and it's the trend I'd bet on lasting longer than Ghibli. The prompt looks like: "Recreate this photo as if it was taken on a disposable film camera in 1996, slightly overexposed flash, grainy film texture, timestamp in orange in the bottom corner." It's popular because it makes a phone photo from last weekend look like it's from a shoebox in your mum's loft.

3. The toy box / action figure family edit

You've probably seen this one: a whole family reimagined as a boxed toy set, complete with plastic packaging, a cardboard header, and little accessory icons for each person. The prompt is usually long and specific ("make each family member into a 3.75 inch action figure in blister packaging, add accessories that match their personality, include a logo at the top that says [family name]"). It's fun, but it's also the one most likely to turn a family of five into a family of four with someone's face blended onto the wrong body.

4. The Renaissance oil painting portrait

"Turn this into an oil painting in the style of a 17th century Dutch family portrait, dark background, formal clothing, visible brushstrokes" is the exact phrasing I've seen copied across dozens of posts. It flatters older family photos especially well because the painterly texture hides the low resolution of scanned prints.

5. The "bring them back" restoration edit

This is the one that gets people emotional, and it's the one I'd be most careful with. People upload an old, damaged photo of a parent or grandparent who has died, and ask ChatGPT to restore the colour, fix the tears and creases, and sometimes to merge that person into a recent family photo so they "appear" in a gathering they never attended. It's touching when it works. It's also the trend most likely to produce something that looks subtly wrong around the eyes, because the model is guessing at detail that no longer exists in the original.

What I did with my own family photos

My sister sent round a Christmas photo from 2003 in our family group chat back in October, one of those flash-lit ones where everyone's eyes are half-closed and the tree lights are blown out white. I ran the restoration prompt on it first, just "restore this photo, fix the colour balance, remove the creases and scratches, keep everything else exactly as it is." That one came out well, closer to how I remember the room looking.

Then I got greedy and tried the disposable camera prompt on a photo of my son and his partner from a barbecue this summer. Three attempts later I still hadn't got one where their dog, who was sitting between them, kept his own markings, it kept turning him into a plain golden labrador he isn't. Small detail, but it's the kind of thing that makes a "just for fun" edit annoying rather than charming once you notice it.

The lesson from both: single-subject or two-person edits are reliable. Group photos with more than three or four people, or photos with pets in the frame, are where these prompts start improvising in ways you don't want.

Step by step: how to get a usable result

• Step 1: Start with the highest resolution version of the photo you have, not a screenshot of a screenshot. Blurry input photos produce blurrier faces, every time.
• Step 2: Upload the photo and describe what should stay the same before you describe the style change. "Keep the same poses, expressions and number of people" at the start of the prompt matters more than the style words that come after it.
• Step 3: Do one style edit at a time. Don't ask ChatGPT to restore colour, remove creases, and turn it into a painting in the same message. Each extra instruction gives the model another chance to drift from the original.
• Step 4: If it's a group of more than three people, check every single face before you share it. I check for asymmetric eyes, extra or missing fingers, and clothing that's changed colour between people who were wearing matching outfits in the original.
• Step 5: Keep the original file. If the edit disappoints you, you want to be able to start again rather than editing an edit, which compounds every small error.
• Step 6: Save both versions with clear file names before you post anything, because these threads move fast and you'll forget which prompt produced which result within a week.

If you want a bigger stash of exact wording to start from rather than writing your own, I've put together 60 copy-and-paste prompts pulled from what's been trending on Instagram, and a separate guide specifically on building family photo prompts with AI tools that goes into more detail on wording for group shots specifically.

Before you upload every photo in your camera roll

Here's the bit that tends to get skipped in the "10 trending prompts to try today" posts. When you upload a photo of your children to ChatGPT, that image goes to OpenAI's servers, and unless you've gone into your settings and turned off "improve the model for everyone," it can be used to train future versions of the model. Most parents doing the disposable-camera trend with photos of their kids have never opened that settings menu. I'm not saying don't do it. I'm saying know that before you do it, rather than finding out afterwards.

The other thing that gets glossed over: a good chunk of the "amazing ChatGPT family photo transformation" posts you see racking up likes aren't edits at all, they're full regenerations. The model isn't adjusting your photo, it's building a new image loosely inspired by it, which is why the family dog changes breed, why a birthmark disappears, why a sibling's glasses vanish. Calling it an "edit" is generous. Managing your expectations around that difference will save you a lot of frustration on the fourth attempt at getting your mother's actual face back.

Why these particular trends, and not others

Trends on ChatGPT move roughly in line with what's cheap to generate and easy to copy from a screenshot. Ghibli-style stuck around because it's one clear instruction. The disposable camera look is having its moment because it works on almost any photo without needing a group at all, which is rarer than you'd think in this space. If you want to see what's replacing these over the coming weeks, it's worth checking what's trending across ChatGPT image prompts this month rather than assuming today's list will still be accurate in a fortnight. These things turn over fast, usually every three to six weeks.

And if family photos aren't really your thing but you want to see what else people are doing with ChatGPT for fun rather than for a client or a business, I've tested a wider set in 47 fun ChatGPT prompts I've tried for personal use, several of which started as photo edits and turned into something else entirely.

What I'd tell a friend to try this week

Start small. Pick one photo, ideally one with two or three people, not the whole extended family at once. Use the restoration prompt if the photo has sentimental damage worth fixing, or the disposable camera prompt if you just want something to post. Skip the toy box edit unless you've got patience for four or five attempts, because it's the most complicated instruction on this list and the most likely to need redoing. And check the small details before you share it. It takes ten seconds and it's the difference between a lovely edit and one where your dad has six fingers on his left hand.

If you want the full set, start at the ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps. To write your own, try the free prompt builder.

Related: chatgpt (this site).

Related: the pets page.

Frequently asked questions

Do I need ChatGPT Plus to do family photo edits like this?

No. Image generation and editing through ChatGPT is available on the free tier, though free accounts get fewer generations before hitting a limit, and paid accounts (Plus, at around £20 or $20 a month) get faster processing and more attempts per day, which matters if you're redoing a group photo four or five times to get the faces right.

How many people can be in a family photo before the edit starts going wrong?

In my testing, two or three people is the sweet spot for reliable results. Once you're past four or five people, especially in a full-body group shot, faces start blending features from the wrong person, and clothing details start swapping between people who were dressed similarly.

Is it safe to upload photos of my children to ChatGPT?

It's generally fine for personal use, but go into your ChatGPT data controls first and switch off "improve the model for everyone" if you don't want images of your kids potentially used for training. Most people doing these trends skip that step entirely without realising it's there.

Why does my edited photo not look like an edit at all, more like a different photo?

Because in most cases it isn't a true edit, it's a full regeneration guided by your original image. That's why small details like a pet's colouring, a birthmark, or glasses sometimes disappear or change, the model is recreating the scene rather than adjusting pixels within your existing photo.

Where to check the details

• OpenAI's own platform documentation