"After the anger comes guilt, the guilt of having exposed your child to a child sex offender." For nearly six months, Pauline, a young mother living near Bordeaux, France, has been asking herself the same question over and over: could her son A., now seven months old, be among the children whose photographs have been used to create AI-generated child sexual abuse material? Like her, several hundred parents in Pauillac, a small town in the Gironde department of around 5,000 inhabitants, and across neighbouring communities, share the same fears.

Pauline is one of the parents whose children were photographed by a professional photographer whom they all knew well from the neighbourhood. Michel A. has been under formal investigation since 26 November 2025 for the possession and distribution of child sexual abuse material, including content generated using artificial intelligence.

Pauline, still deeply shocked, recalls the day she learned about Michel A.’s investigation: "My heart stopped when I read the news on social media.” Shortly after her son's birth, she and her partner had arranged a photo shoot to capture their family’s first moments together. "He gave nothing away”, says the young woman, who never suspected anything at the time. "I could never have imagined that this kind of content could be created using artificial intelligence", she says, still struggling to process what happened as she speaks over the phone.

The photographer was a familiar figure to the community. He worked at local weddings, but also other public events throughout the region, including galas and childrens’ end-of-year performances. "I don't know if you can imagine how many children that represents," Karine, a mother of two daughters who are now 20 and 23 years old, recalls. "He mainly photographed my younger daughter at performances, from the age of seven or eight onwards and for many years afterwards”.

Neither Pauline nor Karine know whether their children's photographs have been used by the photographer to create AI-generated CSAM. The investigation could take months at best, or years at worst. In an effort to find answers, a group of parents, including Karine, set up a Facebook group to pool information and share the limited details they had managed to obtain. Today, the group has over 320 members. "Some people lost sleep over it. Others were deeply worried or furious, and probably still are," Karine recalls.

A contaminated hard drive

The Bordeaux Public Prosecutor’s office identified the 64-year-old photographer, who had previously worked as an accountant for many years. In an email statement, prosecutors explain that equipment seized during the search and examination of the suspect's home and photography studio revealed "the presence of CSAM on a massive scale, at least several thousand files." So far, however, investigators have been unable to accurately identify the minors in the images, further compounding parents’ anguish.

In an attempt to identify the children whose photographs may have been used to generate the material, the Gironde gendarmerie issued a public appeal, asking families to submit photographs of their children, along with their identities, the ages at which they may have been photographed by the suspect, and the relevant locations.

Investigators acknowledged receipt of the information Pauline submitted in January 2026. Since then, there has been nothing but silence and uncertainty. Karine submitted three photographs of her daughters, but wonders whether she identified the correct time period. They date back to when one of her daughters was seven or eight years old. But as the photographer attended all of her daughter’s dance recitals for years, he might just as well have captured her as a teenager to generate the abuse pictures.

For now, Karine can only hope she never receives a call from the gendarmerie telling her that photos of her children match child sexual abuse material found on the suspect's electronic devices.

The suspect has been remanded in custody. For the families caught up in this staggering and unprecedented case in France, there is still no certainty. While they wait, "my daughters prefer to focus on their studies and aren't really sure whether they even want to know," Karine says. Because once the image exists, there is no going back.

Unlearned harm, new tactics

Whereas the first AI-generated videos produced "children with seven or eight fingers," explains Catherine McShane of the Internet Watch Foundation, a British NGO that tracks child sexual abuse material online, today’s outputs are so convincing that analysts increasingly struggle to distinguish genuine images from AI-generated ones. The same challenge confronts officers across cybercrime units involved in such investigations: investigators, police officers, and specialists working for child online safety organizations.

Staff at the Paris offices of the Observatoire e-Enfance/3018 know this only too well. The organization joined the proceedings as a civil party as soon as the Pauillac case became public. Every day, around fifteen helpline advisers provide support by phone and online to parents as well as minors who contact them. They work in shifts of up to 14 hours, from 9 a.m. to 11 p.m.

In 2025, the Observatoire e-Enfance, which is dedicated to combating harassment and digital violence against minors, handled an average of 500 requests a day. These covered issues ranging from cyberbullying and identity theft to child sexual abuse and blackmail involving intimate videos.

Last year, almost a quarter of all reports concerned so-called "Fisha" accounts — social media accounts used to shame, or harass people by posting their intimate content without their permission —, sextortion, or the non-consensual distribution of sexual content. Nearly 3,330 requests for the removal of accounts or online content involved minors and several cases concerned sexually explicit images and videos created with AI-powered tools. Most involved images of digitally undressed victims or deepfakes inserted into sexually explicit videos.

At the time, however, such cases were still relatively rare, and AI-generated content was not tracked separately from other CSAM cases. By mid-2026, the number of such cases had surged, and they are now a distinct reporting category, according to helpline advisers.

Sextortion cases follow a modus operandi of their own: "In exchange for not distributing these AI-generated images or videos, perpetrators demand money, genuine intimate images, or even seek to coerce the minor into a real-life meeting", Véronique Béchu, Director of e-Enfance, explains. In at least one case handled by e-Enfance, this tactic was used to force a minor into sexual exploitation. The victim, who had never shared any sexually explicit images of herself, was sent AI-generated child sexual abuse images depicting her. In exchange for withholding them, the blackmailer demanded that she produce genuine intimate images. But instead of satisfying him, her compliance prompted ever more demands, culminating in demands that she performed sexual acts. The victim was trapped in a cycle of coercion.

Whereas blackmail once relied on intimate images voluntarily shared by victims, today virtually any photograph or video of a child can be transformed into child sexual abuse material. This leaves children vulnerable, even when they have never voluntarily interacted with offenders.

During a recent awareness-raising workshop led by Véronique Béchu for secondary school headteachers, many described what had become an increasingly well-known situation that had become endemic: school disciplinary proceedings are now inundated with cases involving AI-generated sexual images shared, primarily, between students. As early as spring 2024, several cases made headlines in French secondary schools, from incidents reported in Alpes-de-Haute-Provence, Ille-et-Vilaine, and from Haute-Garonne or Ain.

No one to trust

In the small Normandy town of Saint-Hilaire-du-Harcouët, in the Manche department, Marie-Bernard Boudant was among the first school heads to confront the issue.

In March 2025, photographs of 13 teenage girls from the town were manipulated using artificial intelligence to create child sexual abuse images and videos. Around ten of them were pupils aged between 12 and 15 at the private Immaculée Conception secondary school. "As soon as the case was in the hands of the gendarmerie, our main concern was to have the videos removed as quickly as possible so that the girls would not be repeatedly exposed to them. Because you can imagine how quickly this sort of material can spread in a school playground," Boudant recalls.

Some of the girls lost sleep over the ordeal and required long-term support. Others simply wished for their normal lives to return and for the episode to be forgotten. Yet they all shared the same need: for it to be publicly recognized that they bore no responsibility for the videos.

The media flocked to the gates of the small Normandy school to report on the case. "The girls whose images had been stolen and turned into something like this experienced that attention as particularly painful. On top of that, they felt they were being singled out as somehow responsible for what had happened," the headteacher remembers.

The gendarmerie investigation lasted several weeks. Parents, meanwhile, were left feeling powerless and vulnerable. How exactly were their children using social media? What consequences could follow? How should they strike the right balance between parental supervision and protecting their children's image? In an effort to answer these questions, the school organized an information evening in the mid-spring. A legal adviser outlined the procedures to follow in cases involving stolen or manipulated images, but it was not enough for parents who had never experienced such a clash with AI manipulation.

Then, in the middle of the meeting, news broke that the individual responsible for the manipulated images, a 20-year-old man, had been formally placed under investigation. Addressing parents, the gendarme leading the investigation explained: “It is quite remarkable that we managed to identify the perpetrator. We might never have found him, leaving the victims without any criminal justice response and therefore without formal recognition of their status as victims.” Often, he admitted, because the perpetrators are impossible to identify, the safest way to avoid such risks is not to share any personal photos or videos online, no matter how normal such sharing has become.

In this specific case, the perpetrator had acted out of revenge against one of the victims, whom he accused of rejecting him. In an apparent attempt to divert suspicion, he also used photographs of around ten of her friends.

When investigators searched his computer equipment, they discovered CSAM, including videos depicting the rape of minors. He received a two-year suspended sentence and was ordered to pay €1,000 in damages to each victim. The sentence fell short of the maximum three years in prison and €75,000 fine allowed under Article 226-8-1 of the French Criminal Code.

Meanwhile, victims live under the constant threat that screenshots or screen recordings have been made and stored on other devices. "One of the first things our advisers explain is that once content has been generated and published online, it can never be completely erased," says Véronique Béchu.

This is one of the defining characteristics of online sexual violence, which Béchu describes as "amplified violence." A former police commander and child sexual abuse specialist, she adds: "Before the digital age, everything happened behind closed doors. Today, victims live with the constant fear that someone may have seen these images, recognized them, and that the images can resurface at any time."

This investigation was conducted in the framework of our Algorithmic Accountability Reporting Fellowship by our reporting fellow Cécile Debarge.