The Senate of the Republic is analyzing a bill to amend the Federal Copyright Law (LFDA) to regulate AI use in dubbing and voice-over. The proposal, introduced by Senator Beatriz Mojica Morga, expands moral and economic rights so performers and authors can oppose voice cloning without consent and establishes a “notice & takedown” mechanism to swiftly disable deepfakes.

What would change in the law?

• Article 21 (Moral rights): add the ability to oppose modifications, derivative uses, and the cloning or use of works in AI algorithms.
• Article 27 (Economic rights): power to authorize or prohibit (VIII) derivatives via AI, (IX) synthetic reproduction/voice cloning, and (X) use of works for AI training.
• Article 113 Bis (new): requires prior, express, and verifiable consent, proportional remuneration, and clear public notices when an interpretation is fully or partially synthesized by AI.
• Article 114 Nonies (new): penalizes the circumvention/suppression of rights-management information regarding AI use in synthetic interpretations.
• Article 117 (performers): reinforces the right to oppose deformation or cloning of the original voice.
• Article 216 Bis: as a basis for a minimum remedy of 40% of the sale price when there is commercial exploitation that violates rights.

Why now?

The debate intensified after allegations that the voice of actor Pepe Lavat was cloned with AI in an INE spot, along with mobilizations by actors and announcers demanding rules to prevent impersonation. In addition, Mexico’s Supreme Court recently held that works generated exclusively by AI cannot be protected by copyright, setting the legal backdrop.

How it aligns with international trends?

The initiative cites comparative references:

• ELVIS Act (Tennessee, USA), effective July 1, 2024, protecting against unauthorized cloning of voice and likeness.
• UNESCO MONDIACULT 2025 Agenda (Barcelona, Sept 29–Oct 1), focused on culture & AI and the ethical governance of generative systems.

Practical Implications

For studios, brands, and platforms

• Verifiable consent & contracts: annexes specifying AI uses, scope, duration, and compensation.
• Public transparency: visible labels when a voice is synthesized (fully or partially).
• Rapid takedown: internal procedures to remove deepfakes upon valid notice and preserve evidence.
• Model training: avoid using works/performances in datasets without authorization.
• Risks & penalties: damages with a minimum threshold of 40% under Article 216 Bis when exploitation occurs.

For performers and dubbing houses

• Anti-cloning clauses and voice-specific licenses (with registration and traceability).
• Usage labels (e.g., “voice partially synthesized with AI”).
• Evidence strategy: keep demos, timbre samples, and contracts to prove voice identity in potential claims.

Status as of the publication date: the bill was introduced and referred to committees; debate continues in the Senate.