AI Regulation Tracker / National law
Mexico's copyright reform requires written consent to use a person's voice or image in AI systems
Mexico amended its Federal Copyright Law to treat voice as part of a person's image and to require express prior written authorization, with remuneration, before that image or voice is used in any AI system. The change took effect May 15, 2026, and reaches any firm producing synthetic media that touches Mexican performers or individuals.
Mexico has made written, paid consent a precondition for using a real person's voice or image inside an artificial-intelligence system. A decree reforming the Ley Federal del Derecho de Autor, the Federal Copyright Law, was published in the Diario Oficial de la Federación on May 14, 2026, and took effect the next day, May 15. It arrived alongside a companion reform to the Ley Federal del Trabajo, the Federal Labor Law, and the Chamber of Deputies approved the package on April 7, 2026. For any business that generates synthetic media touching Mexican individuals or performers, the compliance clock has already started.
What the reform changes
The central move is a redefinition. The reform replaces the older "retrato," or portrait, concept with "image, including voice," so a person's voice is now treated as part of the protected image right rather than a separate or lesser interest. As reported, using that image or voice "in any format, model or AI system," or publishing outputs generated by such a system, requires express, free, and informed prior written authorization. Per counsel analysis, remuneration is not an optional add-on. It is treated as a condition of the authorization's validity, meaning an unpaid consent may be no consent at all. The exact amended article numbers are not pinned in the public summaries, so firms should read the decree text and confirm citations with local counsel rather than rely on secondary paraphrases.
Who has to change how they work
The reach is broad by design. Baker McKenzie's summary lists broadcasters, commercial voice-over artists, and voice actors as expressly protected, and points to advertising, marketing, media, and entertainment companies, dubbing houses, digital platforms, radio and television, and firms that build AI platforms or train models to create synthetic characters as the parties whose obligations change. A studio cloning a Mexican actor's voice, an agency generating a synthetic spokesperson, or a platform training on identifiable Mexican likenesses now needs a signed instrument that names the person, states the scope, and sets the payment.
What it does not do
The law is not a blanket prohibition on synthetic media. Secondary analyses describe limited exceptions for parody, satire, and creative imitation, provided there is no professional substitution of the artist and no risk of confusion. The right can also be revoked by the person who granted it, which means a one-time release may not hold indefinitely. What the reform does not offer is a safe harbor for training or generating on Mexican voices and faces without documented, compensated permission. Absence of a clear article citation in press coverage is not absence of the duty.
The cross-border angle for US firms
A US advertiser, studio, or AI vendor that produces or publishes synthetic media exploited in Mexico or involving protected Mexican individuals or performers, or that operates through a Mexican entity, falls within this regime regardless of where the model runs. The practical effect is that consent and remuneration paperwork now has to travel with the asset. Secondary reporting describes fines of roughly 1,000 to 5,000 UMAs per infraction, about 5,600 to 28,000 US dollars, doubling for repeat offenses, which counsel should confirm against the decree before relying on any figure. For firms already building consent workflows for US state right-of-publicity and likeness laws, Mexico raises the bar by tying validity to payment.
The operating standard now
The workable assumption is simple. Treat any Mexican voice or likeness that an AI system could reproduce as usable only after an express, free, informed, written, and paid authorization is in hand, scoped to the exact use. Contracts for performers should spell out AI terms and the compensation attached to them. Because the statute is new and the article-level detail is still being parsed, the conservative path is to verify each provision against the DOF text with Mexican counsel before shipping synthetic media into or out of the country.
Frequently Asked Questions
What did Mexico change in its Federal Copyright Law?
A decree published in the Diario Oficial de la Federación on May 14, 2026, and effective May 15, 2026, replaced the older "portrait" concept with "image, including voice." As reported, using a person's image or voice in any format, model, or AI system, or publishing outputs generated that way, now requires express prior written authorization, and remuneration is treated as a validity condition of that authorization.
Who is affected by the reform?
Advertisers and agencies, studios, dubbing and voice-over producers, voice-clone and deepfake developers, AI platforms that generate or train on synthetic characters, media and entertainment companies, and brand and IP counsel operating in or targeting Mexico. The protected parties include performers, broadcasters, voice actors, and individuals whose voice or image could be reproduced.
Does this ban all AI voice clones and deepfakes in Mexico?
No. Secondary analyses describe limited exceptions for parody, satire, and creative imitation where there is no professional substitution of the artist and no risk of confusion. Outside those cases, identifiable clones or simulations require prior written authorization, and the person who granted it can revoke it.
Does it apply to a US company producing synthetic media?
It can. A US advertiser, studio, or AI vendor whose output reaches Mexican performers or consumers, or which operates through a Mexican entity, falls within the regime regardless of where the model runs. The consent and remuneration documentation should travel with the asset.
What is the single most important action to take now?
Obtain express, free, and informed prior written authorization, scoped to the exact use and stating the remuneration, before generating or publishing any AI voice clone or synthetic likeness that identifies a Mexican person, because per counsel analysis payment is a condition of the authorization's validity.
Sponsored Training
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.