Chile's pending deepfake bill would, if enacted, recognize a right over your face, body and voice against AI
By Anthony Guerriero, Founder, The Leveraged Years. SANTIAGO. Published 25 July 2026. Last verified: 2026-07-25.
Most deepfake laws move through criminal codes. Chile is trying something quieter and, for anyone tracking where AI liability is heading, more interesting. A bill introduced by ten members of the Chamber of Deputies wants to hand every person a property-style right over any realistic AI copy of their image, body or voice, and, if passed, would require platforms to take that copy down inside three days. It would call the new protection integridad digital, digital integrity, and it is worth understanding precisely because it is still a proposal you can read line by line at the Congress record rather than a finished statute.
The instrument is Boletin 17.795-19, titled Regula la creacion y difusion de imitaciones digitales realistas de la imagen, cuerpo o voz de las personas, generadas mediante inteligencia artificial (Regulates the creation and distribution of realistic digital imitations of a person's image, body or voice, generated by artificial intelligence). It was filed as a mocion on 21 August 2025 and, per the official tramitacion, remains in the first constitutional stage in its chamber of origin. This is a pending bill, not law.
Key Facts
- Instrument
- Boletin 17.795-19, a mocion (member-sponsored bill), "Regula la creacion y difusion de imitaciones digitales realistas de la imagen, cuerpo o voz de las personas, generadas mediante inteligencia artificial". Source: Senado tramitacion record.
- Issuer
- Ten deputies of the Camara de Diputados de Chile, led author Lorena Fries Monleon, with Andres Giordano, Emilia Schneider, Gael Yeomans and others. Source: authors list, Congress record.
- Filed
- 21 August 2025. Chamber of origin: Camara de Diputados. Source: Senado ficha.
- Who is covered
- Any natural person, national or foreign, living or deceased, whose image, body or voice is reproduced by AI without authorization, plus platforms hosting such content in or from Chile (Articles 2 and 9).
- Consequence (as proposed)
- The bill would set a fine of 100 to 1,000 UTM, damages, and forced takedown, without prejudice to existing criminal actions (Article 10). It would not itself be a criminal statute.
- Status
- Pending. First constitutional stage, general discussion in the Chamber of Deputies; latest recorded floor action 13 July 2026, left pending. Source: tramitacion record.
What Boletin 17.795-19 actually proposes
Strip the bill to its spine and it proposes four things. It would define an imitacion digital realista, a realistic AI reproduction of a person's voice, face, body or expression, and separately define manipulated content that makes someone appear to say or do what they did not (Article 3). It would declare an exclusive right to authorize any such reproduction (Article 4). It would prohibit generating, distributing or even storing these imitations without consent, with narrow carve-outs (Articles 5 and 6). And it would put duties on the platforms that carry the content (Article 9).
The proposed consent standard is demanding. The bill would define consent as prior, express, informed, verifiable and revocable at any time. That last word matters: as drafted, a person could pull authorization back after granting it. For deceased people the right would run 50 years past death and would be exercised by heirs (Article 7), and there is a specific clause for artists and public figures whose performances or speeches are replicated (Article 8).
The "integridad digital" right and the consent rule
Title II of the bill is headed Del derecho a la integridad digital. This is the conceptual move that separates Chile's draft from the transparency-and-labeling approach common elsewhere. Instead of asking whether AI content is disclosed, it asks whether the person depicted said yes. Article 4 reads:
Toda persona tiene el derecho exclusivo a autorizar la reproduccion digital de su imagen, cuerpo o voz mediante IA. El uso no autorizado constituye una infraccion a la integridad digital.
In English: "Every person has the exclusive right to authorize the digital reproduction of their image, body or voice through AI. Unauthorized use constitutes an infringement of digital integrity." The drafters ground the right in Article 19 of Chile's Constitution, numbers 4 and 5 on private life and honor, and cite a Valparaiso appeals court ruling recognizing a person's right over their own bodily image.
There are limits built into the draft. As proposed, Article 6 would exempt content that is satirical, critical or informative, is manifestly identifiable as unreal, and causes neither disinformation nor serious harm. So a labeled parody clip would sit outside the prohibition; a convincing fake presented as real would not.
Where the bill stands, and what happens next
The Congress record is unambiguous that this is early-stage. After filing, the bill went to the Chamber's Comision de Futuro, Ciencias, Tecnologia, Conocimiento e Innovacion in September 2025. That committee delivered its first report on 6 April 2026. The bill reached the floor for general discussion, where the record shows it was left pending on 6 July 2026 and again on 13 July 2026. The executive has attached only simple urgency and filed a formal indication in January 2026.
In plain terms: the Chamber has not approved it. General discussion, the vote on the bill in principle, was still open at the last recorded session. If the Chamber approves it in general and then in particular, the bill would move to the Senate for a second constitutional stage, and only after both chambers agree and the President promulgates it would any of this become enforceable. Anyone claiming Chile has "passed" a deepfake law is ahead of the record.
How Chile's approach compares with Guatemala and Brazil
Three Latin American legislatures are circling deepfakes from different doctrinal starting points. The table below sorts them by mechanism. Only the Chile row is verified at primary source in this piece; treat the peer rows as orientation and check each at its own congressional record.
| Jurisdiction | Instrument | Core mechanism | Status as tracked |
|---|---|---|---|
| Chile | Boletin 17.795-19 | Proposed: civil and administrative "digital integrity" right; consent to reproduce image, body or voice; 72-hour platform takedown; fines 100 to 1,000 UTM | Pending; general discussion, first constitutional stage, Chamber of Deputies |
| Guatemala | Iniciativa 6657 | Criminal-law approach, proposing to punish malicious deepfakes as an offence | Pending initiative in the Congreso de la Republica |
| Brazil | PL 2338/2023 | Broad AI framework with transparency and labeling duties for synthetic content, addressing deepfakes as one category | Approved by the Federal Senate in 2024; under review in the Chamber of Deputies |
The contrast is the story. Guatemala reaches for the penal code, Brazil for a horizontal AI statute built on disclosure, and Chile for a private right the person would personally own and could license. A production using an actor's AI-generated likeness in Chile would not merely need to label it; under this bill it would need the person's revocable consent.
What this bill does NOT do
Precision matters here, because the framing is easy to overstate.
- It would not create a new crime. Article 10 would set fines and damages and would expressly preserve existing criminal actions for identity theft, defamation or fraud, rather than defining a deepfake offence in the penal code.
- It would not ban AI likeness generation outright. Consent, and the satire, criticism and information exceptions in Article 6, would keep clearly-marked and non-harmful uses lawful.
- It would not regulate AI models, training data or developers. The target of the draft is the non-consented realistic imitation and the platforms that carry it, not the underlying technology.
- It does not bind anyone yet. As a pending mocion in general discussion, it imposes no duties, no takedown clock and no fines today.
- It would not replace Chile's data-protection regime. Article 12 would layer on top of Ley N 19.628, adding a consent requirement rather than displacing it.
Why this obscure boletin is worth citing
Chile is not a jurisdiction most English-language AI coverage watches, which is exactly why the primary source is uncrowded. The bill gives you a clean, quotable example of the "right over your likeness" model, with the actual article text, an official title, named sponsors and a live tramitacion trail you can point to. If you are mapping how governments will handle synthetic media, this is the civil-rights branch of the tree, and it is documented at the source rather than paraphrased.
For the full jurisdiction tracker see our AI regulation news hub, the Latin America section, and the related Chile entries.
Primary sources: Senado de Chile, tramitacion of Boletin 17.795-19; full text of the mocion (Congress document record).
Last verified: 2026-07-25.