Roberto Vannacci, the controversial leader of the Futuro Nazionale party, has published an AI-generated deepfake video depicting himself as a Roman general petitioning President Sergio Mattarella for the pardon of convicted jeweler Mario Roggero. The video, posted to Vannacci's Instagram account in late July 2026, presents itself as political theater and raises immediate questions about Italy's recently enacted law criminalizing non-consensual use of public officials' likenesses in synthetic media.
Why This Matters
• Potential legal exposure: Italy's Law 132/2025 mandates up to 5 years imprisonment for distributing non-consensual deepfakes of public authorities without victim complaint. The Vannacci video could potentially fall within this scope, though no charges have been announced.
• Escalating political tensions: The video tests boundaries in Italy's evolving digital political landscape, arriving amid broader concerns about AI-generated content and electoral discourse.
• Roggero case revival: The stunt amplifies public pressure on the Italian Presidency regarding clemency for a jeweler serving 14 years, 9 months for shooting two fleeing robbers in 2021.
The Deepfake Production
Set to epic orchestral music, the video features Vannacci styled as "Robertus Vannaccius," a Roman commander in ancient military dress, entering a senatorial chamber. A figure resembling President Mattarella sits draped in toga-style robes. Vannacci presents a scroll sealed with red wax bearing the Latin word "Gratia" (grace). The dramatization includes figures styled as members of Futuro Nazionale, standing as the faux-Mattarella weighs the request before a blindfolded statue of Lady Justice.
In the climactic sequence, the word "Gratia" bursts into digital flames on the scroll, signaling approval. Vannacci then appears in a prison cell resembling facilities near Milan, removes chains from a prisoner's wrists, and escorts him into sunlight. The final frames show the prisoner embracing his family amid a cheering crowd while Vannacci observes with satisfaction. The caption reads simply: "Grazia per Roggero. Amen." A link directs viewers to a Change.org petition demanding clemency.
Vannacci explicitly labels the content as AI-generated, presenting it as political advocacy rather than covert disinformation. Nonetheless, the use of Mattarella's likeness without consent places the video within the scope of Law 132/2025, which Italy enacted in October 2025 specifically to combat synthetic media abuses.
The Roggero Case: Self-Defense or Vigilante Justice?
Mario Roggero, a 72-year-old jeweler from Grinzane Cavour in Cuneo province, became a flashpoint in Italy's self-defense debate following events on April 28, 2021. When three men robbed his shop, Roggero pursued them into the street and fired multiple shots from a licensed handgun, killing two and wounding a third. Italy's Court of Cassation upheld his conviction for voluntary double homicide and attempted homicide, determining that the shots were fired after the immediate threat had ended and the attackers were fleeing with stolen goods.
Roggero is currently detained at Bollate Prison near Milan, serving a sentence of 14 years, 9 months. His wife has formally requested a presidential pardon—a prerogative that belongs exclusively to the President of the Republic under Italy's Constitutional Court jurisprudence. The case has polarized public opinion: supporters frame Roggero as a victim defending his livelihood against repeat offenders, while legal scholars emphasize that Italian law does not permit lethal force to recover property once danger has passed.
Legal Framework for AI-Generated Content
Italy's regulatory approach to synthetic media has evolved significantly. Law 132/2025 introduced penalties of 1 to 5 years imprisonment for distributing synthetic images, videos, or voices without the subject's consent if they mislead viewers about authenticity and cause harm. When the target is a public authority, prosecutors can act ex officio—without waiting for a formal complaint.
The European Union's AI Act took effect on August 2, 2026, mandating clear labeling of AI-manipulated content. Vannacci's video carries explicit acknowledgment of its artificial nature through its openly fantastical presentation and social media disclosure, though legal experts note that transparency does not necessarily immunize creators from prosecution under national criminal statutes when consent is absent.
What This Means for Residents
For Italians navigating the intersection of technology, politics, and law, the Vannacci video illuminates several realities:
Deepfake laws are now enforceable. Italy's legislative response to AI manipulation is no longer theoretical. Prosecutors possess statutory tools to pursue cases involving public figures, and questions about enforcement in high-profile incidents involving public officials are likely to arise more frequently.
Political expression meets regulatory uncertainty. Vannacci's use of synthetic media as political advocacy tests unclear boundaries between protected political commentary and prohibited non-consensual deepfakes. How courts interpret these boundaries will shape the future of AI-generated political content in Italy.
The Roggero clemency question remains purely presidential. Regardless of viral campaigns or Change.org petitions, Italy's constitutional architecture places pardon authority exclusively with President Mattarella. The video does not alter the legal process: Roggero's family must rely on the formal clemency procedure, typically involving review by the Ministry of Justice before reaching the Quirinale. Public pressure campaigns, while symbolically potent, have no binding effect on the outcome.
The Broader Synthetic Media Landscape
Vannacci's video arrives as Italy grapples with the democratization of deepfake technology. Consumer-grade AI tools now enable convincing impersonations with minimal technical skill, raising concerns across sectors from banking fraud to judicial evidence integrity. Policymakers and platforms face ongoing questions about how to address non-consensual synthetic content while preserving space for political satire and commentary.
Political parties and civil society organizations continue debating how aggressively to police campaign content. Some argue that clearly-labeled satirical content should enjoy broader protections, while others contend that any non-consensual use of a public official's likeness undermines institutional trust. The Vannacci case will likely serve as a reference point for how Italy navigates these tensions in the coming years.
For now, the deepfake remains online, accumulating views and engagement. Whether it results in legal action, policy adjustments, or simply becomes another chapter in Italy's evolving relationship with AI and political expression remains uncertain. What is clear: questions about the boundaries between political commentary and legal liability in synthetic media will continue shaping Italian digital discourse.
