← Conversations archive
VIDEO — 20 Oct 2025

Protecting children from grooming, sextortion and fake AI content

A Vikend popodne interview on Televizija E about grooming, sextortion, AI tools without filters and the steps a child should take.

Televizija E
Published in
Televizija E
Role
Guest
Watch on YouTube

On Vikend popodne on Televizija E, we talked about how to protect ourselves and our children online, from digital violence and from so-called grooming, when predators pose as someone else to target the youngest. We discussed AI that changes a face and a voice live, sextortion, AI tools without filters, a new protection on the iPhone and the steps a child should take when harassed online.

Young people find their way faster

Young people growing up alongside the virtual world are definitely better than earlier generations at recognising someone who poses as someone else online. They accept new technology much more easily and often show their parents how things work, just as we, the generation that grew up with the arrival of the internet and the first computers, were the family IT people. Young people are also targets of digital violence, and because the internet is there 24/7, that violence carries on at home after school. Still, I am very optimistic that they will cope with new apps and ways of communicating much better than my generation or older ones.

AI can already change a face and a voice live

Asked whether a predator could use AI to appear on a webcam as a girl or a boy, I had to say that unfortunately this is already possible. Tools exist that can completely change a person’s appearance and voice live, and in creepy demonstrations a person turns into a celebrity in real time while the model even follows the lighting. As for the signs of AI-generated content, honestly, I don’t know them, because AI is advancing so fast that even I struggle to keep up.

When OpenAI released Sora 2, the internet was flooded with Sora videos straight away, and it took me about two days to come to terms with them. In some of them the intonation, the lip movements and the natural movements of the hands and head are cloned beyond recognition, and Google soon answered with Veo 3.1. I think the companies developing this are intoxicated by the AI race, and worst of all, regulation cannot keep up with something that develops at the speed of light.

Parents need to create trust at home

With the Centre for Civic Education we published an interactive online handbook, and I believe content like this should be published continuously. More cases of digital violence are coming to light, because people talk about it more and children have started to report them, which is exactly the direction we want to move in. I would focus more on parents, because young people will find their way faster, and parents have to create an atmosphere at home in which a child can tell them that a stranger has contacted them with unpleasant requests, such as sextortion or extortion of money. The sooner a child turns to an adult they trust, the sooner it can be stopped and the perpetrator reported, which prevents further violence against others.

How sextortion works

Sextortion is blackmail with a person’s explicit content, usually to extort money in exchange for not publishing it, and with children the threat is that the whole class or school will see it. What is worse, with AI the blackmailer does not even need real explicit content. Here is how an attacker gets children to send photos:

  1. 01The attacker poses as a teenager, for example as a teenage girl to lure a teenage boy, and creates a fake profile with fake photos, even several on Instagram, to look more convincing.
  2. 02He starts a conversation with a lot of interest, praise and shared topics, then sends a fake selfie, someone else’s or AI-generated, so that the child replies with photos of their own.
  3. 03Once he has enough content, he demands money under time pressure: “Do it now, you have 30 minutes to find the money, or I’ll publish everything.”
  4. 04Shame, anger and time pressure leave the child no time to think, so they take a parent’s credit card or empty their savings to pay, or buy a gift card, which also gets the attacker the money quickly.

The host pointed out that in our society, where many parents are conservative, it is hard for a child to tell them about this. That shame is exactly what stops a child from reacting properly. The right reaction is always to block the contact, save the evidence, such as screenshots of the conversation, and report the profile for harassment to the platform where it happened, whether that is Instagram, Snapchat or TikTok. Then talk to someone you trust about what happened, so that the report can be taken further.

AI tools without filters

Commercial tools such as Google’s Gemini, OpenAI’s ChatGPT, Kling AI and Runway have filters built in, but there is also what is called Dark GPT, clones of these tools with all the filters removed. Open-source models such as DeepSeek or Meta’s Llama can be run on your own servers, and once you remove some filters or combine models, you can do practically anything with them. Access to such models is sold on the dark web, the black market of the internet, for making computer viruses and software for cybercrime, and for generating explicit content with no filters. Journalists in war zones or under censorship also use the dark web to share information safely and anonymously.

The iPhone filter that blurs explicit images

When the host asked whether there is software that protects children from such content, I had a very positive answer. iOS, the iPhone’s operating system, recently got a protection against explicit content that works on the device itself. It can be switched on for your own device too, and I believe it is on by default on a child’s profile, blurring the image as soon as it recognises a naked body on the screen.

It even happened to me. During a FaceTime call from the beach, when my wife showed me our child in the pool, my screen blurred with a warning about explicit content, because I had the option switched on without knowing it. A 2025 survey I read recently found that every second child in Montenegro had been contacted at least once by a stranger with explicit content, and I suspect the other half may simply not have admitted it. I think it covered a little over 500 pupils from seven secondary schools, which is a chilling statistic and the reason a control like this is excellent.

Three steps for a child facing online violence

Finally, the host asked what I would tell children who suffer violence online. The first step, easy to advise and hard to apply, is to try to calm down whenever they feel time pressure, anxiety, fear or shame, because only then can they take the next steps rationally:

  1. 01Block the person who is committing digital violence, in any of the forms we talked about.
  2. 02Save the evidence: screenshots, recordings of calls, even made with another phone, voice messages and any other content that was exchanged.
  3. 03Turn to someone you trust: a friend, any older person, a teacher, the school counsellor or someone from the family, and I hope a parent too.

The Ministry of the Interior also recently released a police app with a surprisingly simple way to report potential criminal acts, so in a few clicks you can report peer violence, digital violence and cybercrime. That pleasantly surprised me, and I hope it will have a positive effect.

More conversations & talks