Spain's conservative People's party (PP) posted a campaign video built with AI deepfakes on 21 October, five weeks before the 29 November general election. Likenesses of people linked to Prime Minister Pedro Sanchez's Socialists appear to confess to wrongdoing in the clip. By Friday it had collected nearly 2 million views, a scale that turned a party post into a national debate about AI in politics.
How the video was made and presented
An AI version of Sanchez's wife, Begona Gomez, talks about receiving a university post because of her husband's position. A deepfake of former Socialist prime minister Jose Luis Rodriguez Zapatero talks about hiding jewellery in a safe. Both passages refer to ongoing investigations in which Gomez and Zapatero deny wrongdoing, while other figures in the video appear to admit to soliciting prostitutes and taking kickbacks.
The PP published the video on X with the words "Si, son zurdos" ("Yes, they are left-handed"), quoting a post in which Sanchez said he was left-handed. An AI label sits at the bottom in tiny, nearly transparent text, according to the Guardian's reporting. Party spokesperson Borja Semper said the video was not meant to deceive anyone and described it as a way to use digital tools to highlight Socialist hypocrisy.
Reaction split along political and expert lines. Digital transformation minister Oscar Lopez called the clip "a colossal lie" in an interview with TVE and warned that viewers unfamiliar with AI could confuse it with reality. On Friday, former Socialist party member Leire Diez, whose likeness appears in the video, filed a court complaint accusing the PP of defamation and digital identity theft and seeking immediate removal.
What the case means for business and regulation
For companies that deploy generative video, voice cloning, or labeling tools, the episode shows how disclosure design becomes a compliance issue. A small, low-contrast AI mark did not settle the dispute over deception, even though most of nine legal experts consulted by El Pais said the video was not illegal because it was marked and read as political satire. Vendors selling media production, moderation, and provenance features will face client questions about label size, placement, and audit trails.
Legal clearance alone does not resolve reputational and procedural exposure. Criminal law professor Juan Antonio Lascuraín separated legality from appropriateness and ethics, while researcher Jose Javier Olivas called the video a red line and said putting words in other people's mouths qualifies as disinformation. Firms should therefore check defamation rules, consent for likeness use, platform policies on synthetic political content, and internal approval steps before publishing, since satire arguments did not prevent a court filing.
The marker to watch is how Spanish courts and platforms handle the Diez complaint and whether election authorities or X tighten synthetic-media rules before 29 November. Polls cited by the Guardian point to a PP win that would still require far-right Vox to form a government, so enforcement decisions in the next five weeks will shape what campaigns and technology providers treat as acceptable practice.
