Don’t Mistake the Real
Studio for an Alibi.

Italy’s data-protection authority found that the processing behind deepfake segments of Enrico Mentana, made for an entertainment program, breached the GDPR. It held that the real context, the studio from which his newscast airs, amplified the risk of disinformation, and it cited the mockery Mentana received on social media as confirming that risk.

Enrico Mentana directs an evening newscast. The decision of the Garante per la protezione dei dati personali, Italy’s data-protection authority, records that his real image was used in segments of Striscia la Notizia, an entertainment program published by R.T.I. Reti Televisive Italiane S.p.A. It records that image as “ripresa peraltro all’interno dello studio televisivo”, filmed inside the studio where he directs that newscast, and altered through dubbing done with tools that left the alteration not clearly perceptible. The Garante’s press release of August 7, 2026, differs: image and voice manipulated with artificial-intelligence systems, statements he never made attributed to him, and his image “collocata virtualmente”, virtually placed, in the studio of the broadcaster he works for.

Mentana’s complaint, as the decision recites it, cites comments left under the program’s social media posts, stupido, ignorante, ubriaco, pagliaccio, rosicone, vigliacco: stupid, ignorant, drunk, clown, sore loser, coward. They show, the complaint contends, that many people were convinced the statements were authentic. R.T.I. contends the reverse. Its letter of April 17, 2025, says no one could genuinely have believed they were watching the real Enrico Mentana, citing the tenor of user comments. Mentana’s reply of April 22, 2025, answers from the receiving end. So many people believed the program had used authentic TG La7 footage, he says, that a large number of users called on him, in our translation, to account for the things he had apparently said. R.T.I.’s brief of October 22, 2025, adds that the negative comments were fewer than those showing awareness that this was not the real TG La7 director, and came from users who comment with extreme ease and pay little attention to substance.

Mentana’s complaint put the insults in the record, and R.T.I.’s brief discounted them. Each side reads them as evidence. They dispute what the insults prove: belief, on Mentana’s side; inattention, on R.T.I.’s. The Garante’s decision answers on different ground. The route includes the setting, the studio from which his newscast airs every day, and one clause about the mockery Mentana himself received on social media.

What R.T.I. says it did

The same letter of April 17, 2025, says the only thing R.T.I. did was have Mentana’s voice in the footage dubbed with a different voice, using specialized software. The letter also contends that the videos would make any viewer, even the most inattentive, suspect the reliability of what was broadcast, because the La7 and Striscia la Notizia logos appeared together and because the statements of the Mentana avatar were hyperbolic. Mentana’s complaint reads the TG La7 logo the other way: it was there, the complaint says, “per rendere ancora più credibile la messa in scena”, to make the staging even more credible. The logo is common ground. What it did is not.

Two pages on the program’s own website carry a label. As archived in November 2024, each title opens with the words “Il deepfake di Mentana”: one captured on November 12, the other, “Il deepfake di Mentana svela il vero volto di De Martino”, on November 8. The pages show what R.T.I. published online, not what a viewer saw on air. Read together, the defense makes realness sound like mitigation: his own image, a different voice, a label on the page.

The circumstance the Garante named

The Garante put the realness on the other side of the scale. Its findings run one observation beside the next: his actual image; dubbing whose alteration was not clearly perceptible; statements plausible at first glance; the setting. The decision describes the real context he works in, the studio from which the news program he directs airs daily, as a circumstance apt to make the processing particularly insidious for his personal identity, amplifying the risk of its becoming a “veicolo di disinformazione per il pubblico”, a vehicle of disinformation for the public.

The same sentence names what confirmed that risk: “come confermato dai numerosi messaggi di scherno e/o di offesa ricevuti via social dal reclamante”, the numerous mocking or insulting messages Mentana received on social media. The decision’s phrase is messages he received; the complaint’s phrase is the comments listed under the program’s posts. We read this as the clause this edition turns on. To R.T.I., the comments showed that no one could have believed it was really Mentana. To the Garante, the mockery that reached the real man confirmed the risk of disinformation.

The disclaimers, the Garante held, were not clear enough, at least for the technological knowledge of “un pubblico medio o comunque poco attento”, an average or not very attentive audience. The context, together with what the deepfake of a well-known journalist said, would have required more explicit communications, made more evident at various points of the broadcast, “in modo da intercettare anche gli utenti meno attenti o quelli che si fossero collegati a programma già iniziato”: so as to reach even the least attentive users, or those who tuned in after the program had begun.

The least attentive viewer

Two phrases sit in the same record. In the decision’s recital of R.T.I.’s letter, the viewer is “anche il più disattento”, and the logos, in that telling, put him on guard. In the Garante’s standard, the viewer is “anche gli utenti meno attenti”, and the broadcast would have had to reach him, late arrival included. One phrase assumes the viewer. The other builds for him.

A ban and a warning

The Garante found that the processing violated the principles of lawfulness, fairness and transparency in Article 5 of the GDPR, and Article 25, data protection by design and by default, for want of adequate technical and organisational measures. Provvedimento n. 577 of July 23, 2026, bans further processing of Mentana’s data “con le modalità sopra descritte”, in the manner described, except storage for possible use in legal proceedings, under Article 58(2)(f). A warning goes to R.T.I. under Article 58(2)(b), and the measures are entered in the Garante’s internal register. The operative part, dated July 23, 2026, contains no administrative fine. The Garante treated the novelty of the legal question as a factor making the warning a proportionate measure.

The real studio was not an alibi. In the decision it is a circumstance, apt to make the processing particularly insidious. The ban targets a manner of processing, not the program. The Garante named the build: a real setting, an alteration not clearly perceptible, and disclaimers not clear enough for an audience the decision called average or not very attentive.

The Agentics read

Agentics reads this case through Digital Derangement Syndrome™ (DDS™): AI-mediated systems failing to recognize legitimate authority. No system in this record fails to recognize Mentana. The characteristic that fits, read in reverse, is Trust Transfer Failure, the fourth clinical characteristic: real-world expertise and earned authority never translated into machine-readable trust. Mentana’s authority was already legible, in his face and his studio, and a machine wore it.

Where the argument stops

Mentana speaks here only through his complaint and his reply, as the decision recites them; R.T.I. speaks through its letter and its brief, recited the same way. This edition did not view the segments themselves. The decision says his image was filmed inside the studio; the Garante’s press release says it was virtually placed there. The press release is the Garante’s summary of the decision, not a second source, and this edition cannot say which description is right. The two R.T.I. pages quoted show what its website carried in November 2024, not what reached a viewer on air. Translations from the Italian are ours.

Beyond the press release of August 7, 2026, this edition searched for no appeal, measure, or enforcement record after July 23, 2026, nor R.T.I.’s thirty-day report. The mapping to DDS™ is an editorial reading, not a finding by the Garante.