Audio version on our Youtube channel: Here

In the previous article about sharenting (which can be read HERE), we outlined the risks associated with sharing children’s images online - from identity theft and deepfake manipulation to cyberbullying and the use of personal data within a broader digital ecosystem. While many of these threats may seem distant or difficult to imagine, in reality they are increasingly taking on very concrete, personal forms.

One tangible example of how real the consequences of sharing children’s data online can be is the campaign “A Message from Ella” created by Deutsche Telekom. We encourage you to watch the video HERE.

In this short video, a computer-generated version of Ella speaks to her parents from a cinema screen. Based on just a single photograph, the creators were able to generate a digital version of the girl, several years older. Ella addresses her parents directly, explaining the real consequences of sharing photos and information about their child on social media.

The message is striking. What appears to be an innocent act of sharing everyday moments can lead to outcomes far beyond what parents anticipate. What they consider a harmless memory becomes data in the digital ecosystem. Data that can be analyzed, combined, and repurposed in ways entirely detached from its original meaning or intent.

It is important to remember, however, that this is not just a fictional campaign scenario.

According to data cited in educational materials and reports on sharenting:

• over 80% of children have a digital footprint before the age of two

• up to 75% of parents share information about their children online

• a significant proportion of parents do not restrict access to this content, meaning it is visible far beyond close family circles (often over 70–80% of viewers are unrelated individuals or complete strangers)

This means that a vast amount of sensitive information reaches people we are unable to verify. And such information can be used for identity theft, tracking and profiling, or the creation of deepfakes. And these situations are not hypothetical.

Real-World Cases

Deepfake and ransom

In 2023, a widely reported case in the United States involved Jennifer DeStefano, who received a phone call in which she heard her 15-year-old daughter crying. The girl was screaming that something terrible had happened and that she needed help. Moments later, a man joined the call, claiming he had kidnapped her and initially demanding a ransom of one million dollars.

As the mother later explained, she had no doubt it was her daughter. The voice, the emotions, the way she spoke - everything matched.

In reality, her daughter was completely safe. The call was a scam based on AI voice cloning, created from audio material available on social media.

Similar cases are becoming increasingly common. In one instance, a woman in Florida transferred $15,000 after speaking with what she believed was her daughter, who had never actually called. Scammers used artificial intelligence to replicate her voice based on publicly available recordings.

Since then, the development of AI technologies has accelerated significantly. What once required specialized tools and expertise is now accessible to almost anyone. A short voice sample and basic information about family relationships shared online are often enough.

“Adoption role-play” accounts

A less obvious but well-documented phenomenon related to children’s online presence is the rise of so-called baby role-play accounts.

These are fake profiles created on platforms like Instagram or TikTok, using real children’s photos taken from parents’ accounts. The mechanism is simple, but the consequences are very real.

Publicly available photos shared by parents are downloaded by strangers and used to create entirely new profiles. In these profiles, the child is given a new name, age, personality, and even a fictional life story. The child becomes a character in someone else’s narratives, often within so-called adoption role-play, where users interact with the account by commenting, “caring” for the child, or acting out family relationships in the comments.

In one case described by Fast Company, a mother from North Carolina discovered that photos of her daughter had been taken and used in exactly this way. Someone was posting her child’s images under a different name, inventing stories and details about her life and health.

Despite attempts to contact the account owner and report it, nothing changed. Her daughter’s images continue to circulate online, stripped of their original context and beyond her control.

Materiały CSAM

Even more concerning conclusions come from reports by organizations such as the Internet Watch Foundation and Europol.

They indicate that publicly available photos of children shared by parents can become source material for the creation and distribution of sexually exploitative content within closed online environments. In many cases analyzed, there was no hacking or data breach. The material came from legal, publicly accessible social media posts, later downloaded and processed outside the parents’ control.

For example, in 2025, Daniel Joseph Broadway was sentenced to 78 months in prison for possessing CSAM, including content generated using artificial intelligence. Investigators found over 30,000 AI-generated images and videos of this nature on his devices. He used generative tools not only to create new material but also to transform existing photos of real individuals found online into sexualized content.

The court emphasized that such material is persistent and repeatedly consumed, meaning each use constitutes a renewed violation of the victims’ privacy and dignity.

In 2024, the Internet Watch Foundation reported a 380% increase in cases involving AI-generated abuse content. That year marked a shift from static deepfakes to the emergence of realistic AI-generated video circulating on the dark web.

Consequences That Are Hard to Ignore

What connects all these cases is the source of the data they rely on. Photos, videos, and information shared as part of everyday parental activity cease to function solely as private memories. They enter a digital circulation where they can be downloaded, processed, and used in contexts completely detached from their original purpose.

At the same time, technological development means that just a few seconds of publicly available material are now enough to create a realistic replica of someone’s voice or image. The tools required are widely accessible and inexpensive.

This fundamentally changes the nature of the risk. Control over where an image ends up, and what can be done with it, is effectively lost.

Sharenting, therefore, is no longer just about sharing family life. It becomes part of a broader data ecosystem in which the boundary between the personal and the public is increasingly blurred. And while most of this content is shared in good faith, with innocent intentions, its digital life begins to operate independently beyond any meaningful control.