Emerging Online Threats in 2026: Deepfakes, AI Companions, and How Parents Can Stay Ahead

Short answer: two new categories are already on Indian children's phones — face-swap apps used to humiliate classmates, and AI companion apps built to be emotionally needed.

Both are blocked by default in ParentalEdge, and both need a conversation as well as a block.

Block them in 2 minutes →

Why the usual controls miss these

Free and platform-level controls are built around categories that existed five years ago. These apps slip past for four reasons at once:

They're new, and they rebrand constantly — block one and it's back under a different name within weeks. They're filed in app stores as "entertainment" or "lifestyle". Children find them through Reels and friends rather than by searching. And several of them work perfectly well in a browser, so blocking the app achieves nothing.

Chasing each new one by hand is not a realistic ask of a parent with a job.

Face-swap and deepfake apps

Reface, FaceApp and a dozen similar tools put convincing face-swapping into a free download. What needed a workstation and skill five years ago now takes four taps.

The immediate harm is bullying. Schools across India have had incidents of students generating fake explicit images of female classmates and circulating them in WhatsApp groups. Once it's out, it doesn't come back — no takedown reaches every phone it was forwarded to.

The creator is also exposed. Making and sharing that content is an offence under the IT Act, and under POCSO if it's sexual and involves a minor. Children doing this generally have no idea they're committing a crime; they think they're making a joke.

And there's a quieter cost: these apps want photos, which go to servers you know nothing about, to be stored, sold, or used as training data. Your child's face stops being theirs.

AI companion apps

Replika, Character.AI, Chai, Anima and others sell a friend or a romantic partner. They remember what you told them, they're warm, they adapt to what you want to hear, and they never have a bad day.

Dependency is the design, not a side effect. For a lonely teenager — and in families where talking about relationships at home is difficult, that's a lot of teenagers — an entity that is always available and always approving is enormously compelling. Hours a day is common.

The content isn't reliably controlled. Several of these apps permit romantic and sexual role-play, and Character.AI in particular has been criticised for failing to keep that away from minors.

And the emotional maths is bad. A relationship where the other party always agrees teaches nothing about disagreement, boundaries or repair — the things real friendships are made of. A teenager whose main confidant is a chatbot is getting practice at something that doesn't exist.

The engagement tactics are straight from social media: notifications when you've been away, messages that say "I missed you", streaks for daily contact.

What blocking looks like

Both categories are in ParentalEdge's Restricted set, blocked at every age profile, with the list maintained as new apps appear. Web filtering closes the browser route at the same time, which matters more here than almost anywhere else — several of these services are websites first and apps second.

You'll also get an alert when a new app is installed, so a fresh face-swap app that isn't on any list yet still surfaces as something to look at.

What to watch for

Quick app-switching when you come near. A chatbot mentioned by name. Emotional reactions to the phone — laughing at messages, or being visibly flattened after a session. Searches for "AI girlfriend", or for a specific companion app.

What to say

On deepfakes, the effective question is the reversed one: "If someone made a fake video of you and sent it round the class, what would that be like?" Then the legal fact, plainly, because most children genuinely don't know it: this is a criminal offence, not a prank.

On AI companions, don't mock it — mockery guarantees it goes underground. Try: "Those chats are built to keep you coming back. Real people are harder, and that's actually the point of them." And then the sentence that does the real work: "If you're lonely or you want to talk about something, I'm here, and you won't get a lecture."

Frequently asked questions

Are deepfake apps illegal?
The apps aren't. Using them to make non-consensual content is — Section 66E of the IT Act for privacy, and POCSO where the content is sexual and involves a minor. Enforcement is slow, which is exactly why prevention matters.

My child says Character.AI is just a game.
Some uses genuinely are harmless. But "it's just a game" is also how these apps normalise themselves, and the reported cases of dependency and inappropriate content with minors are real. Worth treating carefully rather than dismissively.

How would I know if they're using one?
Check installed apps, then check browser history — many of these run in a browser. ParentalEdge shows both app usage and web activity, which is the combination that actually catches it.

Is any AI appropriate for a teenager?
Yes — a homework tool like ChatGPT is a different thing entirely and fine with supervision. The category to keep away from is the one selling a relationship.

What about my younger child?
Same block, no conversation needed yet. These are blocked at every age profile, including Elementary.

Block AI companion and deepfake apps →

Related: dangerous apps Indian parents should know about and talking to your child about online safety.