Deepfakes in the classroom: homes raided in Entre Ríos over AI-altered intimate images of female classmates

By El Litoral · October 26, 2026.

Police in Entre Ríos (Argentina) carried out three raids on Saturday in Villa Adela, Puerto Yeruá and Calabacilla, in the Concordia area, and seized 11 mobile phones. The case concerns the alleged AI manipulation of photographs of three students at the Escuela Agrotécnica General San Martín to make them appear topless.

According to the newspaper, the accused are reportedly at least three students from the same school. The operations were carried out by officers from the Calabacilla police station, under the direction of Gender prosecutor Agustina Solé and by order of the Family and Juvenile Court. The phones will now undergo forensic examination to determine whether they contain relevant material and how the images were generated and shared.

The incident reportedly took place on Friday morning. According to accounts gathered by El Litoral, the students and their families went to the school to complain, there were tense moments at the entrance and as the school buses departed, and police drew up a report and invited them to file a formal complaint. The outlet itself stresses that the acts attributed to the students are part of an investigation and must be established in the proceedings: they are alleged and unproven.

The case is not isolated. The same outlet links to another recent case in Ituzaingó, where four students are under investigation for creating fake nude images of classmates, and to a piece on a UNICEF figure: one in five adolescents has reportedly suffered digital sexual abuse. That statistic is cited in El Litoral's headline and we have not verified it.

**Our take.** This is the real face of AI's short-term harm, and it should be looked at without drama or excuses. No superintelligence is needed: a photo "undressing" app, a profile picture and a messaging group are enough. The technical barrier has dropped to zero and the harm falls on the girls, with psychological and reputational consequences that no subsequent removal fully reverses.

There is also an institutional mismatch. Schools, families and juvenile justice operate at a pace that technology has already left behind. The fact that intervention came after a public complaint by families rather than through a pre-existing protocol suggests the system reacts rather than prevents. The alleged perpetrators are also minors, so the response must combine accountability and education, not just punishment.

Broadly, solutions lie on three fronts: providers of image-editing tools blocking the generation of nudes from photos of real people, platforms swiftly removing this material, and digital and relationships education entering classrooms with the same weight as other subjects. Regulating out of panic does not work; targeting this specific use, which is demonstrable and harmful, does.

It does not change our core thesis: AI can bring enormous advances in health and abundance in the long run, but that transition will only be legitimate if we protect those who today bear its worst effects. Meanwhile, cases like this measure the maturity of our institutions better than any benchmark.