Spain approves national AI law to align with EU AI Act, introducing new bans and obligations

Why it matters

Spain’s government approved a draft law on May 26 to govern artificial intelligence under national rules aligned with the European Union’s AI Act. The bill, approved by the Council of Ministers, would ban a set of AI practices including sexual deepfakes and social scoring, impose fines of up to 35 million euros, and require disclosure when some content has been generated or manipulated by AI. The government said the aim was to ensure AI is used in a “trustworthy, ethical and rights-based” way. ### What exactly did Spain approve? The Council of Ministers approved a project law — not a final enacted statute — called a bill for the good use and governance of artificial intelligence. La Moncloa, the Spanish government’s official site, said the measure was adopted on May 26, 2026, and Infobae, citing government information and EFE, reported that it now moves to the Cortes Generales for parliamentary debate and approval. (lamoncloa.gob.es) The Spanish measure is designed to adapt domestic law to Regulation (EU) 2024/1689, the EU AI Act, which created a harmonized legal framework for AI across the bloc. EUR-Lex describes that regulation as setting rules for the development, placing on the market, putting into service and use of AI systems in the European Union. ### Which AI practices would be banned? (lamoncloa.gob.es) The draft bans AI systems used to generate sexual deepfakes or realistic sexual content identifying real people, according to EFE’s summary of the text. It also bans social scoring — using AI to rate people based on social behavior or personal traits for decisions such as access to subsidies or credit — along with manipulative subliminal techniques, exploitation of vulnerabilities tied to age or disability, certain biometric categorization, emotion-recognition uses, and predictive surveillance practices. (eur-lex.europa.eu) Those categories broadly track the EU AI Act’s treatment of prohibited practices. EUR-Lex says the AI Act bars certain uses considered to create unacceptable risk, while also imposing separate obligations on other categories such as high-risk systems. ### What would companies and public bodies have to do? The draft requires AI-generated content to be labeled. (infobae.com) EFE reported that images created and distributed with AI would have to include watermarks marked with the initials “IA,” and the government-backed reporting said users must be able to distinguish artificial content from reality. (eur-lex.europa.eu) The bill also keeps stricter obligations for systems classified as high risk rather than banning them outright. Infobae said those systems would face added controls, and the Spanish government described the law as setting rules to assure reliable and guaranteed use. The EU AI Act separately contains transparency duties for AI-generated or manipulated image, audio and video content, including disclosure that content was artificially created or manipulated. (infobae.com) ### How large are the penalties? The Spanish draft sets fines ranging from 6,000 euros to 35 million euros depending on the seriousness of the violation, according to Infobae. The highest sanctions would apply to conduct such as marketing or putting prohibited tools into service. (infobae.com) The 35 million-euro ceiling matches the top figure highlighted across Spanish reports on the bill approved Tuesday. EFE’s account said the draft creates a sanctioning regime with multimillion-euro penalties tied to the prohibited and high-risk categories laid out in the text. (infobae.com) ### Why does this matter beyond Spain? Spain’s move comes as EU member states begin translating the bloc’s AI framework into domestic enforcement and administrative machinery. The EU AI Act is already in force at the European level, and Spain’s bill sets out national bans, disclosure rules and sanctions that would sit alongside that framework. (noticiasdealava.eus) The next formal step is parliamentary processing in the Cortes Generales. Infobae reported that the government had previously approved a first anteproyecto in March 2025, submitted it for public consultation, and revised it after reviews including by the General Council of the Judiciary before returning it to the cabinet on May 26, 2026. (infobae.com) (eur-lex.europa.eu)

Key numbers

• Spain’s cabinet approved an AI governance bill on May 26 that would align national rules with the EU AI Act and ban several uses.
• The draft sets fines from 6,000 euros to 35 million euros, bans sexual deepfakes and social scoring, and requires labels on AI-made content.
• Spain’s government approved a draft law on May 26 to govern artificial intelligence under national rules aligned with the European Union’s AI Act.
• The bill, approved by the Council of Ministers, would ban a set of AI practices including sexual deepfakes and social scoring, impose fines of up to 35 million euros, and require disclosure when some content has been generated or manipulated by AI.

What happens next

• Spain’s government approved a draft law on May 26 to govern artificial intelligence under national rules aligned with the European Union’s AI Act.
• The government said the aim was to ensure AI is used in a “trustworthy, ethical and rights-based” way.
• La Moncloa, the Spanish government’s official site, said the measure was adopted on May 26, 2026, and Infobae, citing government information and EFE, reported that it now moves to the Cortes Generales for parliamentary debate and approval.

Sources

• lamoncloa.gob.es
• eur-lex.europa.eu
• infobae.com
• infobae.com
• noticiasdealava.eus

Quick answers