Spain's PP deepfake ad shows a tiny AI label won't protect voters from synthetic smears

Spain's conservative People's Party (PP) has published a campaign video on its official social media accounts. It uses artificial intelligence to make realistic copies of people linked to Prime Minister Pedro Sánchez's Socialist party and shows them appearing to admit wrongdoing. According to The Guardian, by Friday the video had nearly 2 million views. The general election is next month.

The Guardian describes the details. An AI-generated version of Sánchez's wife, Begoña Gómez, talks about getting a university post thanks to her husband's job. A deepfake of former Socialist prime minister José Luis Rodríguez Zapatero talks about hiding jewellery in a safe. Both lines refer to ongoing investigations, and both Gómez and Zapatero deny any wrongdoing. Other figures appear to confess to soliciting prostitutes and taking kickbacks. The only disclosure is a note at the bottom of the screen, in what the paper calls "tiny, nearly transparent font", saying the video was made with AI.

The government responded quickly. Óscar López, Spain's digital transformation minister, told the broadcaster TVE that the party was "telling people a colossal lie." He warned that anyone not used to AI could easily mistake the video for reality. Leire Díez, a former Socialist party member whose likeness appears in the video, has filed a court complaint. According to reports, it accuses the PP of defamation and digital identity theft and asks for the ad to be taken down at once. None of these accusations has been tested in court.

The legal picture is less clear than the outrage. Most of the nine legal experts consulted by El País said the video was probably not illegal, because it was labelled as AI-generated and its tone was political satire. "Everyone knows those people didn't say those things," said Íñigo Navarro, a former law dean at Comillas Pontifical University. Juan Antonio Lascuraín, a criminal law professor at the Autonomous University of Madrid, drew the key distinction: legality is one question, and whether the ad is appropriate or ethical is another. José Javier Olivas, who researches disinformation and AI biases at Spain's National University for Distance Education (UNED), called the video a "red line" for how campaigns are normally run. The debate is not only Spanish: AI-made political ads are also a talking point ahead of next month's US midterms.

Our reading: this case exposes the weak spot in the main tool regulators rely on, which is labelling. A label in nearly invisible type meets the letter of the rules and defeats their purpose. The legal argument assumes "everyone knows" the confessions are fake. But people see the video in a fast-scrolling feed, and the fake confessions repeat real allegations from real investigations. Satire works by exaggerating what someone actually says. This ad does something else: it makes up admissions on topics already in the courts. That is closer to fabricated evidence than to caricature. The worry is also not mainly about technology. The material does not say which tool was used, and that hardly matters. It was a major party, on its official accounts, deciding that synthetic likenesses of rivals are fair campaign material. If that becomes normal, the next step is predictable: the other side answers in kind, and in a polarised country the cost falls on trust in any video at all.

Where it goes from here: in the short term, expect pressure for disclosure rules with real standards, such as label size, placement and visibility, and for consent requirements before a real person's likeness is used in a political ad. These are small, enforceable changes, not a ban on AI in campaigns. In the long term we remain optimistic. The same tools can make campaigns cheaper and more accessible, translate political messages and help voters check claims. That only holds if the rule becomes "don't put words in real people's mouths", and it needs to be enforced rather than buried in the fine print.