AI Kiss Privacy and Safety: What Happens to Your Photos

Uploading a photograph of your own face to a generative model is a reasonable thing to hesitate over. This page answers the questions people actually ask before doing it, using what the Privacy Policy and Terms genuinely say rather than reassuring language that promises more than the product delivers.

Where the policy is specific, this page quotes it. Where the policy is not specific, this page says so instead of filling the gap.

What happens to a photo after you upload it

The sequence is short.

Your images are prepared in your own browser first. They are resized down and, in two-photo mode, composed into a single frame: two halves of one canvas, one person in each. That combined image is what actually gets uploaded, not your original files sitting untouched somewhere.

The upload goes to cloud storage. A generation job is queued against it. A model reads that starting frame plus a short instruction describing the motion, and produces the video. The finished MP4 goes back into storage so your gallery can show it to you, and you download it from there.

There is no human review step in that loop. If you want the mechanics in more depth, the explainer on how these clips are made covers the model side.

What is stored, and what the policy says

Here is the honest version.

The combined image you upload and the video you get back are held in cloud storage. They have to be, or the app could not hand you a result minutes later or show you a gallery when you come back.

The Privacy Policy states that photos are processed securely and are not stored permanently unless you save them, and that temporary processing files are automatically deleted after generation. It also says the service does not sell, trade or rent personal information to third parties, that data is encrypted in transit and at rest, and that your photos are not used for any purpose other than generating the videos you asked for.

What the policy does not do is publish a specific retention window in days. So this page will not invent one. If you want something gone at a particular moment, the reliable route is to ask rather than to wait.

The full text is on the privacy page, and the rules about what you may upload and create are on the terms page. Both are worth a read before you upload someone else's face.

How to ask for deletion

Write to privacy@aikiss.app, the contact address on the Privacy Policy, or to support@aikiss.app.

The web app signs you in anonymously, which is good for privacy and slightly awkward for support: there is no email address on file to match a request against. Make the request easy to act on by including the approximate date and time you created the videos, the device and browser you used, and anything else that narrows it down. The policy says the team typically responds within 48 hours.

The policy also lists the rights people have depending on where they live, including access, correction, deletion, restriction of processing, portability and withdrawal of consent. Those requests go to the same address.

One practical note if you are deleting on someone else's behalf. If you made a clip of a friend or a relative and they have since asked you to take it down, delete your own copies first, remove the post from wherever you published it, and then send the deletion request. The post is the part the world can see, and it is the part that stops mattering the moment you remove it.

No account, and what that means for you

There is no sign-up, no email and no password. The app creates an anonymous session in your browser and everything hangs off that.

The upside is real: nothing to leak, no password to reuse, no marketing list to end up on, and no profile building up over time.

The trade-off is just as real. Your gallery belongs to that browser on that device. Clear your site data, switch to a different phone, or open the app in a private window, and previous videos will not be there. Download anything you care about as soon as it finishes. This is the single most common surprise people report, and it is the one thing you can prevent entirely in ten seconds.

The app overview covers the practical side of running the whole thing from a phone browser.

Consent is the rule that matters

Every other safety consideration is downstream of this one.

Use photos of yourself. Use photos of people who have agreed to being in an AI-generated kiss video. That is the whole rule.

It excludes: strangers whose photos you found online, public figures and celebrities, anyone who would be surprised or upset to see the result, and anyone under eighteen in a romantic clip. The Terms prohibit content that violates a person's privacy or publicity rights and content depicting minors inappropriately, and access to the service is for people aged thirteen and over, with parental consent required under eighteen.

Publishing is a second, separate permission. Someone can happily watch a clip you made of the two of you and still not want it on a public account. Ask about the post as well as the generation, especially for family photographs where other people appear.

Deepfakes, and where the line sits

The underlying technology is related to what people mean by deepfakes, so it is worth being clear about what separates the two.

The difference is not the model. It is consent and intent. A clip of you and your partner, made from your own photographs, labelled as an AI edit, deceives nobody and harms nobody. A clip of a person who did not agree, presented as though it were real, is a different act entirely regardless of how it was produced.

Three questions cover almost every case. Would the person in the photo be pleased to see it? Would a viewer understand that it is generated? Would you be comfortable if the subject saw exactly how you had described it? If any answer is no, the video should not exist.

It is also worth being careful with clips of people who have died. Animating a photograph of a grandparent can be a genuinely moving thing to make, and it can also land badly with relatives who were not asked. Treat the wider family as part of the consent question rather than an audience for the finished post.

What the platforms do with AI content

Every large platform now labels realistic AI-generated media, either from the metadata attached to the file, from creator disclosure, or from its own detection.

That means the label is arriving with or without your cooperation. Adding a line yourself is simply better: it comes across as confidence, it heads off the accusation in the comments, and it prevents the odd situation where the platform contradicts a post that implied the footage was real.

Platforms also remove non-consensual intimate imagery quickly and permanently, and increasingly remove the accounts behind it. There is no version of the misuse case that ends well.

A short checklist

Before you upload anything:

1. Are these photos of you, or of people who said yes?
2. Have you asked about posting as well as about making it?
3. Is anyone in the frame under eighteen or a public figure? If so, stop.
4. Have you downloaded previous videos you want to keep?
5. Do you know where to write if you want something deleted?

Get those right and the rest of this is a small, pleasant thing to make. The photo requirements guide covers the technical side of choosing images once you are comfortable with the ethical side, and the app is where the actual work happens.