They walked free after accepting responsibility for fabricating stories with AI
En San Luis Potosí, dos periodistas —madre e hija— salieron de prisión tras aceptar responsabilidad por haber usado inteligencia artificial para fabricar relatos que vinculaban a funcionarios estatales con el crimen organizado. Su liberación, enmarcada en la llamada Ley Serrano de 2025, abre un capítulo inédito en la historia del periodismo mexicano: el momento en que la tecnología que prometía amplificar la verdad comenzó a ser juzgada por fabricarla. El caso no cierra con su salida de La Pila; las investigaciones continúan, y con ellas, la pregunta de hasta dónde puede extenderse una ley diseñada para combatir la desinformación antes de rozar la libertad de prensa.
- Eréndira Reyes y su hija Alejandra Hermosillo pasaron más de dos semanas detenidas en La Pila acusadas de usar IA para crear contenido falso que implicaba al gobernador y a la Guardia Civil Estatal con el Cártel Jalisco Nueva Generación.
- La Ley Serrano, aprobada apenas en noviembre de 2025, se convierte en el primer instrumento legal aplicado en México para sancionar penalmente el uso no autorizado de narrativas generadas por inteligencia artificial.
- Ambas periodistas aceptaron su responsabilidad y accedieron a una suspensión condicional del proceso, evitando el juicio a cambio de un año de conducta supervisada, residencia fija y reparación del daño causado.
- Las autoridades advierten que las investigaciones se extienden a otros comunicadores que habrían empleado tácticas similares de manipulación con IA, señalando que este caso es apenas el primero de una campaña de aplicación más amplia.
- El trasfondo inquieta a defensores de la prensa: una ley concebida contra los deepfakes se aplica ahora a periodistas que publicaron sobre funcionarios públicos, difuminando la frontera entre desinformación ilegal y reportaje erróneo pero protegido.
Eréndira Reyes, directora de los sitios San Wicho Times y Despertar 3.0, y su hija Alejandra Hermosillo abandonaron el penal de La Pila en San Luis Potosí el sábado pasado, después de más de dos semanas de detención. Habían sido arrestadas el 21 de mayo acusadas de publicar historias falsas generadas con inteligencia artificial que vinculaban al gobernador Ricardo Gallardo Cardona y a la Guardia Civil Estatal con el Cártel Jalisco Nueva Generación.
El fundamento legal del caso es la Ley Serrano, promulgada en noviembre de 2025, que penaliza la creación y distribución de imágenes, voces o narrativas generadas por IA sin autorización cuando dañan la reputación o identidad de personas. Un juez de control aprobó la suspensión condicional del proceso penal tras confirmar que ambas mujeres reconocieron su responsabilidad y manifestaron disposición para cumplir las condiciones impuestas.
Durante el próximo año, Reyes y Hermosillo deberán reparar el daño causado por sus publicaciones, mantener un domicilio fijo, presentarse periódicamente ante las autoridades y demostrar buena conducta. Los detalles exactos de la reparación —si implica compensación económica, rectificaciones públicas u otras medidas— no fueron precisados por la fiscalía.
Más allá del destino individual de las dos periodistas, el caso proyecta una sombra más larga. La fiscalía estatal confirmó que las investigaciones continúan contra otros comunicadores sospechosos de haber usado herramientas de IA con fines similares, lo que sugiere una estrategia de aplicación sistemática de la nueva ley. Para quienes defienden la libertad de prensa, la pregunta urgente es si un marco legal diseñado para combatir los deepfakes puede convertirse en un instrumento para silenciar el periodismo crítico, incluso cuando ese periodismo resulta ser falso.
Two journalists walked out of a state prison in San Luis Potosí on Saturday after spending more than two weeks behind bars. Eréndira Reyes, who runs the news websites San Wicho Times and Despertar 3.0, and her daughter Alejandra Hermosillo had been held since May 21st at the state detention facility known as La Pila. They were released after accepting responsibility for publishing false stories generated with artificial intelligence—a decision that allowed them to avoid trial and return home under strict conditions.
The case hinges on a law passed in November 2025 called the Ley Serrano, which makes it illegal to create or distribute AI-generated images, voices, or narratives without authorization. The law was designed to punish deepfakes and synthetic media that damage people's reputations or identities. In this instance, prosecutors determined that Reyes and Hermosillo had used AI to fabricate visual and narrative content that falsely implicated the state governor, Ricardo Gallardo Cardona, and the Estatal Civil Guard in connections to the Jalisco New Generation Cartel, one of Mexico's most violent criminal organizations.
A control judge approved what's known as a conditional suspension of the criminal process, effectively erasing the preventive detention order that had kept both women in custody. The state prosecutor's office confirmed that both defendants acknowledged their responsibility and expressed willingness to comply with the judicial conditions imposed on them. The agreement represents a middle path between full prosecution and complete dismissal—a mechanism increasingly used in Mexican criminal justice to resolve cases without trial.
To maintain their freedom, Reyes and Hermosillo must satisfy several requirements over the next year. They must repair the damage caused by their publications, maintain a fixed residence, appear for periodic check-ins with authorities, and demonstrate good conduct throughout the probationary period. The specifics of what damage reparation entails—whether financial compensation, public retractions, or other measures—were not detailed in the prosecutor's statement.
What makes this case significant extends beyond these two individuals. Prosecutors indicated that investigations remain active against other media professionals suspected of using similar AI manipulation tactics. The state's attorney general's office signaled that the same legal framework being applied to Reyes and Hermosillo would be used to pursue additional communicators and journalists. This suggests a broader enforcement effort to police how news organizations and independent media figures use artificial intelligence tools, particularly when those tools are deployed to spread false narratives about public officials or state institutions.
The case sits at the intersection of two competing concerns in modern Mexico: the genuine problem of disinformation campaigns that can incite violence or undermine public trust, and the potential for such laws to be weaponized against critical journalism or political speech. The Ley Serrano was framed as a protection against deepfakes and synthetic media fraud. Yet its application here—to journalists publishing stories about government officials and security forces—raises questions about where regulators draw the line between illegal manipulation and protected reporting, even when that reporting turns out to be false.
Citações Notáveis
Both defendants acknowledged their responsibility and expressed willingness to comply with the judicial conditions imposed on them— San Luis Potosí State Prosecutor's Office
A Conversa do Hearth Outra perspectiva sobre a história
Why did they accept responsibility so quickly? Two weeks in prison seems like a short time to reverse course entirely.
The conditional suspension offer was likely attractive because it avoided a full trial with uncertain outcomes. They could walk free immediately rather than face months or years of proceedings.
But doesn't accepting responsibility mean they're admitting the AI stories were deliberately false? Or could they have believed what they published?
The prosecutor's office said they used AI to create false visual and narrative structures. Whether they believed the content or knowingly fabricated it for impact—that distinction isn't clear from what was released.
The governor and the cartel connection—that's a serious accusation. Why would they publish something like that?
That's the unanswered question. It could have been for clicks, for political motivation, or because they genuinely believed it. But once they used AI to generate it rather than report it, they crossed into the territory the law was designed to catch.
Are other journalists worried about this law now?
They should be. If the state can prosecute media figures for AI-generated false content, the definition of what counts as 'false' and what counts as 'AI-generated' becomes a tool of control. The law sounds reasonable until it's used against reporting the government doesn't like.
So this could chill actual journalism?
That's the risk. When journalists see colleagues detained for publishing false stories—even if those stories were genuinely false—it changes what people are willing to investigate and report.