Google Play age ratings on AI apps: why a 3+ can mean very little
Upload two photos, get a short video of the two people kissing. On 15 October, MediaNama found an app that does roughly that, AI Kiss & Hug Video Maker from a Surat developer, still listed on Google Play in India with an age rating of 3+. That was the week India’s child rights commission, the NCPCR, summoned Google over the whole category.
The obvious reaction is that somebody at Google made a mistake. The more useful one is to ask how a rating like that gets made at all, because once you know, you stop trusting the number and start using the settings that actually do the work.
Nobody at the store watched the app before it got its number
Google’s own developer help is plain about the process. To get a rating, the developer “must fill out a rating questionnaire on the Play Console”, and the app is then given ratings “based on your questionnaire responses”. The questionnaire belongs to the International Age Rating Coalition (IARC), which describes its own system as one that “assigns ratings instantly without rating authorities’ review of the product prior to its release.”
So a 3+ means the developer answered a list of questions about their app in a way that produced 3+. IARC says its rating authorities monitor ratings afterwards and correct them through the stores, and Google warns developers that “misrepresentation of your app’s content may result in removal or suspension”. Both of those happen after the app is live and downloadable.
Google’s page for users says that “content ratings don’t tell you whether an app is designed for users of a specific age.” The label describes the content the developer declared. It was never a stamp saying a child should be using this.
ThePrint’s investigation in August shows the gap in practice. It found dozens of these apps on Google Play in India, several downloaded more than a million times, several registered to individual developers in Gujarat. One app’s own landing page said it was “intended for users aged 18 and older” while the store rated it 12. Another, KiHu, was rated for ages three and up. Technology lawyer Dhruv Garg told ThePrint that parental controls “only work if the rating is accurate.”
What these apps do with a photo
ThePrint found that one of the most downloaded apps declared in its data safety section that it may share photos, videos and device identifiers with third parties, without naming them. Privacy policy links on several apps led nowhere. Some users said their uploaded pictures stayed online after they closed their accounts. The same app charged Rs 449 a week or Rs 2,999 a year for an “ad-free” mode, with reviewers complaining that cancelling did not stop the charges. And the videos carried no AI label, even though India’s amended IT Rules, in force since 20 February, require synthetic content to be “clearly and prominently labelled”. ThePrint notes it is contested whether those rules reach a standalone generator app at all.
Google’s AI content policy does already ban “AI-generated non-consensual deepfake sexual material” and apps “primarily intended to be sexually gratifying”. Google has removed some apps in this category, and ThePrint reported several that had vanished since early August. But MediaNama’s October check shows one rated 3+ was still there.
For a child, the risk runs in two directions. Their own face can end up on a server nobody can name. And a classmate’s face can end up in a video that classmate never agreed to. Garg’s view, quoted by ThePrint, is that making intimate imagery of someone without consent can be an offence under the IT Act and the Bharatiya Nyaya Sanhita. Most children who try these apps will think of it as a joke. It is worth saying out loud, before it happens, that it is not one.
The three settings that do the work
If the rating is a developer’s self-description, the dependable gate is you. On an Android phone your child uses, through Google’s Family Link:
- Require approval for every download. Open Family Link, choose your child, go to Controls, then Google Play. Under “Purchases & download approvals”, tap “Require approval for” and choose “All content”. Google says free apps trigger the request too, so a free kiss and hug app cannot slip through on price.
- Block anything already installed. Approval and content filters do not touch apps already on the phone. Google says plainly that apps downloaded earlier stay visible “even if they’re outside of the rating you set”. In Family Link, choose your child, go to Screen time, then Time limits, then App limits, pick the app and switch it from Allowed to Blocked. Google says it takes effect in about five minutes.
- Read the data safety section, not the stars. When an approval request arrives, open the listing and scroll to data safety. “May share photos and videos with third parties” on an app whose whole job is faces is your answer.
[[photo:1]]
Keep the approval notifications switched on. Google notes that if you turn off notifications for Family Link or the Play Store, “you won’t get notified when there’s a request”, and the request simply waits.
Rules are coming, slowly
MediaNama reported on 12 November that the NCPCR plans to release draft child safety guidelines for public consultation and finalise them in early 2027. The draft reportedly asks for age assurance that goes beyond self-declaration, private accounts by default and limits on in-app spending. An official called the inputs being sought “exploratory in nature”, which suggests they may not bind anyone.
Our view: that is a reasonable direction and a long wait. In the meantime, the kiss and hug apps are a useful lesson that travels well beyond India. A family in Singapore reading a Play Store listing sees the same kind of self-declared number, and our earlier piece on Singapore’s app store age checks found the same weakness from the other side. If a fake image of your child does appear, the Singapore reporting route is worth knowing in advance.
Tonight, the five-minute version: turn on “All content” approval, scroll through the apps already on your child’s phone, and tell them, in your own words, that a made-up video of a real person is still about a real person.
Mentus AI makes a mentoring app for children, so we have a commercial interest in parents trusting the apps their children use. Nothing in this article depends on our product.
Was this useful?
Sources
- AI 'kiss and hug' app rated for ages 3+ still on Play Store as NCPCR summons Google · MediaNama
- Upload anyone's pic, AI will make intimate video of you. 'Kiss & hug' apps test India's IT rules · ThePrint
- India plans to finalise child social media guidelines in 2027 · MediaNama
- Apps & Games content ratings on Google Play · Google Play Help
- Content Ratings (Play Console Help) · Google
- Understanding Google Play's AI-Generated Content policy · Google
- How IARC works · International Age Rating Coalition
- IARC FAQ · International Age Rating Coalition
- Manage your child's apps on Android · Google For Families Help
- Manage your family's purchases on Google Play with purchase approvals · Google Play Help
- How to set up content restrictions on Google Play · Google Play Help