Predators Only Need 20 Images Of Your Child To Create Deepfakes

We’re used to hearing that the internet never forgets, but according to one online safety expert, it now remembers, learns, and generates – and parents might be unknowingly fuelling it.

Speaking on Today with Claire Byrne, Mick Moran, former Garda and Interpol investigator turned CEO of hotline.ie, warned that even a handful of innocent photos shared online can be twisted by AI into something horrifying. Specifically: child sexual abuse material (CSAM).

MORE: New Study Finds Rising Demand for AI-Generated Child Abuse Images

DISCOVER: Should you be hiding your children’s faces online?

Image @newscomauhq via Pinterest

“He says 20 images is all that’s needed,” Moran said, referencing research by Professor Karsten Maples of the University of Warwick. “That’s eventually going to get down to one, as AI gets stronger.”

It’s the kind of development that makes your stomach drop. Thanks to increasingly multimodal AI tools like OpenAI’s GPT-4 or Google’s Gemini, machines can now process photos, video, audio, and text all at once. They don’t just “see” your pictures – they learn from them. And once your child’s image is online? You’ve effectively lost control over how it’s used.

MORE: Australia Proposes Social Media Ban for Teens Under 16

DISCOVER: TikTok Profiting From Livestreams Involving Children According To BBC Reports

“Fundamentally, any information you share online can be used in ways you never intended,” Moran said.

What’s more, some platforms may treat your public post as a green light. “You would imagine that data protection rules would mean that those images couldn’t be used,” Moran explained, “but if you’re sharing them publicly, that expectation of privacy is gone.” And worse, some platforms interpret this public sharing as “implicit consent” — much like signing a consent form on behalf of your child for a school trip.

The big players like Google and Meta have relatively strict regulations and enforcement, Moran says. The real problem? The under-the-radar companies hoarding data in the shadows. “It’s not twee to say that ‘data is the new gold’. It’s a commodity,” he added.

Image courtesy of rawpixel.com

MORE: Facebook is building a children’s version of Instagram

DISCOVER: ‘Jools Law’ Could Give Parents Access To Their Children’s Social Media Accounts

So, should parents stop sharing photos of their kids entirely? Moran says absolutely not. “It’s important that’s not used to panic parents… Absolutely share your pictures,” he said. The key is sharing them privately – via closed groups or with well-managed privacy settings.

His advice: dive into those settings and lock them down. Too often, people don’t realise that default settings are public, meaning anyone – and anything – can access them.

“All of that should be turned on by default. But it’s not,” he said. “And there’s the trap people are falling into.”

Bottom line? You don’t need to go off-grid – just get privacy-savvy. Because in the age of generative AI, even a toddler’s birthday photo can become part of something no parent ever imagined.