Is AI-Generated Content Legal in Brazil? What the Law Actually Says

If you're generating AI images or video in or about Brazil, it helps to know where the actual legal lines sit rather than guessing. The short version: Brazil doesn't yet have a single enacted law that regulates "AI-generated content" as its own dedicated category. A federal AI bill, PL 2338/2023, passed the Senate in December 2024 and has been sitting with a special committee in the Chamber of Deputies since April 2025, with a final vote repeatedly pushed into 2026 — it is not yet in force. In the meantime, a generated image or video in Brazil is measured against the same general-purpose laws that already cover any manipulated or fabricated media: the Lei Geral de Proteção de Dados (LGPD), Civil Code portrait and personality rights, copyright law, and the Penal Code's provisions on non-consensual sexual imagery. This isn't legal advice — Brazil's framework is genuinely still moving through 2026, and for anything beyond casual personal use you should consult a lawyer licensed in Brazil about your specific situation.

Generate fictional AI content

There's no Brazilian law that regulates "AI-generated content" as its own category yet

As of 2026, Brazil generally treats an AI-generated image or video the way it would treat any other photograph, illustration, or manipulated media — what tends to matter legally is what the content depicts and how it's used, not the fact that a generative model produced it. That puts the laws most likely to actually apply to a generation squarely in existing statutes: the LGPD (Lei nº 13.709/2018) where personal data or a real likeness is processed, the Civil Code's image-rights and personality-rights provisions (direito à imagem, Arts. 11 and 20), Brazil's Copyright Law (Lei nº 9.610/1998), and the Penal Code's non-consensual imagery provisions discussed below.

The LGPD and its regulator, the ANPD, are the closest thing to an active enforcement backstop

The LGPD is Brazil's general data protection law, and it applies to processing personal data — which can include a real person's face, voice, or other biometric-adjacent likeness used to train or steer a generation. In September 2025, Brazil converted the ANPD (Autoridade Nacional de Proteção de Dados) into a fully independent regulatory agency with its own budget and expanded enforcement powers, and its published 2026–2027 priority themes reportedly name AI and emerging technologies as a top enforcement focus, following an earlier ANPD preliminary study on generative AI. None of that is a content-specific rule for adult AI imagery, but it signals that data-protection enforcement around AI is intensifying rather than staying dormant.

Non-consensual sexual imagery of a real, identifiable person is the highest-risk area — with or without AI

Article 218-C of Brazil's Penal Code, added by Lei nº 13.718/2018, makes it a crime — punishable by 4 to 10 years in prison plus a fine (raised from the original 1-to-5-year range by Lei nº 15.280/2025, in force since December 2025) — to offer, share, sell, publish, or otherwise disseminate a sex scene, nudity, or pornographic depiction of a person without that person's consent, with the penalty increased if done for revenge or humiliation or by a former intimate partner. The statute's own exception clause, which carves out anonymized journalistic or scientific use of imagery of someone "over 18 years old," confirms this provision is about general non-consensual disclosure involving adults, not a statute specific to people under 18. Legal commentary and reporting in Brazil describe this article as already being applied by courts to AI-fabricated sexual images of real people, on the reasoning that the harm is the non-consensual depiction and disclosure, not the exact production method. This is the single clearest area of legal exposure for AI content depicting an actual, identifiable adult without consent.

A federal AI bill (PL 2338/2023) is still pending, not yet law

PL 2338/2023 — often described as broadly inspired by the risk-based structure of the EU AI Act — cleared the Federal Senate on December 10, 2024, and was sent to the Chamber of Deputies, where it was referred to a special committee established in April 2025. As of mid-2026 the bill remains under committee review, with reporting describing provisions around risk classification and synthetic-content labeling as still being negotiated, and the ANPD expected (but not yet confirmed by enacted text) to serve as the lead supervisory authority. Separately, a Civil Code reform bill (PL 4/2025) has been reported as proposing a new article addressing the use of a living or deceased person's image or voice via AI. Treat any specific labeling, disclosure, or classification requirement from either bill as provisional until final text is enacted and published.

What this actually means if you generate fictional AI content

For entirely fictional, non-real-person AI characters — content that doesn't depict, name, or closely imitate any actual identifiable individual — the current Brazilian legal picture generally looks lower-risk than content built around a real person's face or voice, since Article 218-C, Civil Code portrait rights, and LGPD likeness-processing concerns are specifically about real, identifiable people. That said, Brazil's Copyright Law still applies a broadly comparable infringement standard to AI output as it does to human-made work, so a generation that closely reproduces a specific existing copyrighted character likely carries similar exposure to a human-made copy. None of this is a substitute for legal advice tailored to what you're actually doing — Brazilian AI, data-protection, and criminal law in this area is genuinely in flux through 2026, and only a lawyer licensed in Brazil can tell you how it applies to your specific situation.

FAQ

Is there a specific Brazilian law about AI-generated content?

Not a single enacted one, as of 2026 — the federal AI bill (PL 2338/2023) passed the Senate in December 2024 but remains under committee review in the Chamber of Deputies. Generated content is currently judged under existing law instead: the LGPD, Civil Code image-rights provisions, Copyright Law, and Penal Code provisions on non-consensual imagery. This is a general summary, not legal advice.

Is it legal to generate an AI image of a real person in Brazil without their consent?

This is the highest-risk category under current Brazilian law as we understand it. Article 218-C of the Penal Code criminalizes distributing a non-consensual sexual or nude depiction of a person, with 4 to 10 years in prison plus a fine (raised from the original 1-to-5-year range by a December 2025 amendment, Lei nº 15.280/2025), and reporting indicates courts have already applied it to AI-fabricated images of real people. Civil Code portrait-rights and LGPD likeness-processing rules can also apply. Consult a lawyer licensed in Brazil before treating this as settled for your specific case.

Does Brazilian law treat fictional AI characters differently from real people?

Generally, yes, in terms of risk category — the criminal and portrait-rights provisions described above are specifically about a real, identifiable person's image or likeness, so entirely fictional, non-real-person AI content sits in a materially lower-risk category. Copyright law can still apply if a generation closely reproduces a specific existing copyrighted character.

Where can I find the official text of Brazil's LGPD?

Brazil's Presidência da República publishes the official LGPD text (Lei nº 13.709/2018) on the Planalto legislative portal — that's a more authoritative primary source than relying on secondhand summaries, including this one.