Spain has given final approval to a new AI law that will fine companies for failing to label synthetic content, a move that will also affect AI providers operating in Catalonia and across the country. The government says the law is designed to align Spanish rules with the EU AI regulation and to make it easier for users to tell real content from virtual content.
Once the law comes into force, companies that create synthetic content with AI and do not label it could face penalties. The EU regulation, which took effect on 2 August, requires AI-created audiovisual content to be identified, but it does not set sanctions. Those fines will come through national laws such as Spain's.
Prime Minister Pedro Sánchez linked the issue to Pope Leo XIV's recent encyclical on AI, saying on X that AI is not neutral and that digital power can lead to new atrocities if it is not directed towards the common good. Digital Transformation Minister Óscar López also backed that view, saying the encyclical protects human rights and contrasting it with the position of major tech oligarchs.
The law applies extraterritorially, so any AI provider offering services in Spain must comply, regardless of where it is based. It also sets out rules on algorithmic transparency, accountability for company executives and protection for minors. The minister said the framework mirrors the EU regulation in banning subliminal techniques used to manipulate decisions without consent, as well as chatbots that target potential gambling addicts or exploit age, disability or socio-economic vulnerability.
It also bans sexual deepfakes and the generation or alteration of child pornography. Non-sexual deepfakes will remain legal, but they must be clearly labelled with the initials AI. The law sets penalties from €6,000 to €35 million for the most serious cases, including the commercialisation of prohibited AI systems.
Public administrations are exempt from fines, although ministry sources said admonition formulas will apply. The Spanish Data Protection Agency will continue to handle biometric data, while AI use affecting justice will fall under the General Council of the Judiciary. Other AI applications will be overseen by Aesia, an agency with around 50 analysts tasked with monitoring compliance. The final text was approved this Tuesday, a year after its initial approval, following public consultation and review by several bodies, including the CGPJ. More Catalonia news