ChatGPT caricature trend: the prompt, how to do it and privacy

The ChatGPT caricature trend is the one that filled social feeds in early February 2026: a cartoon version of you, big head and all, surrounded by the tools of your job. You make it by giving ChatGPT a photo and asking it to draw you based on everything it knows about you. It is fun, and it takes two minutes. It is also a neat demonstration of how much ChatGPT remembers, which is why privacy warnings, including one from Forbes on 9 February 2026, followed within days.

The ChatGPT caricature prompt

TechRadar and The Tab both reported the same wording. Upload your photo, then type:

Create a caricature of me and my job based on everything you know about me.

If the style is off, add a line asking for a classic, funny caricature with exaggerated features and a bright, hand-drawn look, which is the kind of tweak TechRadar suggested. If ChatGPT knows little about you, add the facts yourself:

I'm a primary school teacher in Soweto. I coach netball, I'm always on my
phone with parents on WhatsApp, and I never go anywhere without rooibos.

How to make a ChatGPT caricature, step by step

1. Open ChatGPT in the app or at chatgpt.com and log in, so it can use your chat history and memory.
2. Choose a clear, well-lit photo of your head and shoulders. Crop out anything private in the background, such as your house number, a street sign or your children.
3. Tap the plus or attach button and upload the photo.
4. Type the prompt and send it.
5. Wait. Images take longer than text replies, and OpenAI describes free-plan image generation as slower.
6. Ask for changes in plain words: “make me look less grumpy” or “put a braai in the background”.

Does it work on the free plan?

Yes. According to OpenAI, free accounts can upload images and create images in ChatGPT, but each has its own usage limit, and OpenAI’s pricing page describes free image generation as limited and slower. When you reach a limit, ChatGPT tells you, and you wait for it to reset. Free accounts also get memory, but a more limited version than the paid plans.

The paid Go and Plus plans add more image creation and a longer memory. Our guide to ChatGPT free vs paid in South Africa sets out what each plan costs in rand and what it adds.

How ChatGPT memory makes the caricature work

The trend only works because of memory. OpenAI says that when Memory is on, ChatGPT can remember relevant details from your chats and other sources, which can include past chats, saved memories, custom instructions, files and connected apps, depending on your plan and region. When you ask for a caricature based on everything it knows, it pulls from that store: your job, your pets, your hobbies, even how you described your desk.

That is why different users with the same prompt get very different pictures, and why a brand-new account gets something generic. You can see some of what ChatGPT holds about you by asking it what it remembers, or by opening Memory in Settings.

The privacy catch with ChatGPT caricatures

The picture itself is the smaller risk. The bigger ones are what you hand over and what you then post.

• Your photo and chats go to OpenAI. On personal accounts, OpenAI says it may use chats and remembered information to improve its models while the setting Improve the model for everyone is on. You can switch it off in Data controls.
• The finished caricature can say too much. Security firm Bitdefender warned that these images can reveal workplace details, location clues and children’s identities. A public post showing your job, employer and interests is useful material for a scammer building a convincing message.
• Backgrounds leak. A car registration, a school badge or a name on a door says more than you think.
• Other people’s faces. Uploading someone else’s photo means sharing their face and details without them. Ask first.

POPIA, South Africa’s data protection law, does not apply to purely personal or household activity, so making a caricature of yourself for fun is your own choice. It does apply when a business does it. If a company makes caricatures of staff or customers, it is processing their personal information and needs a lawful basis, usually consent. Children’s information has extra protection: POPIA generally requires the prior consent of a parent or guardian. This is general information, not legal advice.

How to turn off ChatGPT memory

OpenAI’s help centre gives these controls, though what you see depends on your plan and region:

1. Open Settings, then Personalization, then Memory.
2. Switch off Reference saved memories and Reference chat history, or the single Memory switch if that is what your account shows. Turning off saved memories also turns off chat history references.
3. To wipe what it holds, use Delete and turn off memory from the memory summary, or delete items one by one. Deleting a chat does not automatically delete a memory that was saved from it.
4. For a one-off question, start a temporary chat and choose Unpersonalized. Temporary chats do not create or update memories.
5. Separately, switch off Improve the model for everyone under Data controls if you do not want your chats used to train OpenAI’s models.

OpenAI says that when you switch off chat history references, what it remembered from past chats is scheduled for deletion within 30 days, and it may keep logs of deleted saved memories for up to 30 days.

Using the trend at work

Caricatures can make a fun team post or a year-end slide, but treat staff photos like any other personal information: ask each person, explain where the picture will be used, and keep customer details out of the prompt. If your team uses ChatGPT for real work, Zaiq’s guide on how to use ChatGPT for business covers the safe way to set it up.