The Ethics of AI Images: Consent, Other People's Likeness and Disclosure
A practical guide to using AI avatar and image tools responsibly: getting consent, why you shouldn't generate other people's faces, special care with minors, and when to disclose AI images.
· 5 min read
AI image tools make it easy to put a face into any scene, any outfit, any style. Used on yourself, that’s mostly harmless fun. Used on someone else — or shared without context — the same tools can embarrass, mislead or seriously harm people.
This guide isn’t legal advice, and laws differ between countries and change quickly. It’s a practical framework for using AI avatars and images in a way you’d be comfortable defending, built around three ideas: consent, respect for other people’s likeness, and honest disclosure.
Consent: start with your own face
The simplest ethical position is also the one most avatar tools are designed for: make images of yourself. You know what you’re comfortable with, you can delete what you don’t like, and nobody else’s identity is at stake.
When other people are involved, consent should be:
- Informed. They understand what will be generated, in what style, and where it might be shared.
- Specific. Agreeing to a cartoon avatar for a group chat isn’t agreeing to a realistic image posted publicly.
- Revocable. If they change their mind, you delete the images and the photos you uploaded.
Asking takes ten seconds. “Can I upload a few photos of you to make an anime version for our team Discord?” is an easy question. If the answer is no, or you can’t ask, don’t.
Other people’s likeness
Friends, family and partners
Even with good intentions — a birthday gift, a joke for a group chat — uploading someone’s photos to an AI service means sending their face to a third party whose retention and training policies they never agreed to. Our privacy checklist applies to them too. Ask first.
Exes, colleagues and strangers
Generating images of someone who hasn’t agreed, especially someone you have a conflict with, can quickly become harassment. Many services ban it in their terms, and in many places some uses are illegal.
Public figures
Celebrities and politicians are often the subject of parody and commentary, and some of that is protected expression in many legal systems. But realistic images that could be mistaken for genuine photos — a politician at an event that never happened, a celebrity endorsing a product — can mislead people and may infringe “right of publicity” or advertising laws, which in the US vary from state to state. If an image could plausibly be taken as real, don’t make it, or make the fiction obvious.
Children
Be especially cautious with images of minors. Children can’t meaningfully consent to having their likeness processed, and images shared online are hard to retract. Many AI avatar services restrict uploads of minors in their terms. A good rule: don’t upload photos of children who aren’t yours, and think carefully even about your own.
Bright lines
Some uses are not grey areas at all:
- Sexual or intimate imagery of real people without their consent. This causes serious harm and is illegal in a growing number of places. In the US, the TAKE IT DOWN Act, signed in 2025, criminalises publishing non-consensual intimate images, including AI-generated ones, and requires platforms to remove them on request. The UK and other countries have also made creating or sharing such deepfakes an offence.
- Impersonation and fraud. Using someone’s AI likeness to deceive others — for scams, fake profiles, fake endorsements — is unethical and often illegal.
- Fabricated evidence. Images designed to make it look like someone did something they didn’t.
- Sexualised images of minors of any kind, real or generated.
Responsible AI image tools block or prohibit these uses. If a tool doesn’t, that tells you something about it.
Disclosure: when to say it’s AI
Not every AI image needs a label. An obviously stylised anime or pixel-art avatar isn’t going to fool anyone. Disclosure matters more as images become realistic and as the context makes people assume they’re genuine.
A practical scale
| Situation | Disclosure |
|---|---|
| Clearly stylised avatar (anime, pixel, cartoon) on a social or gaming profile | Usually unnecessary |
| Realistic AI headshot on a personal social profile | Be honest if asked; a note in your bio is a nice touch |
| Realistic AI headshot for professional use (press, company pages) | Consider a real photo instead; if not, disclose |
| Realistic AI image of a scene or event shared as news or commentary | Always disclose clearly |
| Advertising or sponsored content using AI images of people | Disclose, and follow platform and advertising rules |
| Any realistic AI image of another real person | Disclose, and have their consent |
Platform and legal rules
Several platforms now ask creators to label realistic AI-generated or altered content. YouTube requires creators to disclose realistic synthetic or altered content in certain cases, TikTok asks for labels on realistic AI-generated content, and Meta applies “AI info” labels on its apps. The EU AI Act also introduces transparency obligations for AI-generated “deepfake” content, phased in over the coming years. The details differ and keep evolving, so check the current rules of wherever you post.
Provenance and watermarks
Some tools embed invisible watermarks or attach content credentials (the C2PA standard) to generated images, recording that they were made with AI. Don’t strip them. They help people understand where an image came from, and removing them may breach a platform’s or tool’s terms.
Respecting artists
Styles are not owned, and “anime” or “watercolor” are broad traditions. But generating images explicitly “in the style of” a specific living artist, then using them commercially, raises fair questions about credit and livelihood, and some platforms restrict it. Choose general style descriptions rather than naming individual artists, and if you love a particular artist’s style, consider commissioning them.
A quick self-check before you share
Before posting an AI image, ask:
- Whose face is this? If it isn’t mine, do I have their clear, specific consent?
- Could someone mistake this for a real photo? If yes, have I made it clear it’s AI?
- Could this embarrass, mislead or hurt anyone? If yes, don’t post it.
- Am I following the platform’s labelling rules?
- Would I be comfortable if the person in the image saw exactly how I’ve used it?
How BecomeAI approaches this
BecomeAI is being designed as an avatar studio for your own face. Our intended terms prohibit uploading other people’s photos without consent, and we plan to make downloads easy to identify as AI-generated. Combined with our intended policy of deleting source photos after generation, the goal is a tool that’s fun to use and hard to misuse. We’ll publish the full terms before launch.
Summary
- Make AI images of yourself by default.
- Get informed, specific consent before using anyone else’s photos — and never use children’s photos casually.
- Avoid realistic fakes of real people; never create intimate or deceptive images.
- Disclose realistic AI images, follow platform labelling rules, and keep provenance data intact.