A Fake of Your Face, a Fake of Your Company: In Saudi Arabia and Jordan, the Law May Not Help

US · CanadaAmérica Latina

Weekly AI Watch

Every week, what changed and what it means for you. No hurry, no alarm.

Week of 23 October 2026EN | ESPast editions18,276 papers found · 17,064 read · 5,349 belonged · 37 stories passed every check

Observatorio de la
Inteligencia Artificial

Cada semana, qué cambió y qué significa para usted. Sin prisa y sin alarma.

Semana del 23 de octubre de 2026EN | ESEdiciones anteriores18,276 estudios encontrados · 17,064 leídos · 5,349 correspondían · 37 notas pasaron todas las comprobaciones

weeklyAI · Week of 23 October 2026weeklyAI · Semana del 23 de octubre de 2026

← Your rights← Sus derechos

analysis of texts · Frontiers in artificial intelligence · la publicación, 2 jul 2026 · gratis

Una ley que solo protege a las personas de carne y hueso

Un jurista saudí propone reformas para que las empresas puedan reclamar por daños de imagen. Son sugerencias, no ley vigente, y ningún tribunal de esos países ha resuelto todavía un caso así.

Versión breve · la versión detallada sigue, unos 7 min

Pregunte al chat de weeklyAI

Pregúnteme por este estudio: a quiénes se estudió, qué encontró y qué no dice.

Las conversaciones se guardan mientras exista weeklyAI, para mejorar la publicación. Se responde en el idioma en que usted escribe.

El estudio, de un vistazo
Quiénes: Las leyes civiles de Arabia Saudita y Jordania sobre daños por videos falsos
Cuántos: No aplica: es un análisis de textos legales, no un conteo de personas
Dónde: Arabia Saudita y Jordania
Cuándo: Leyes de 2023 (Arabia Saudita) y 1976 (Jordania); artículo publicado en julio de 2026
Tipo de estudio: análisis de textos legales (estudio doctrinal)
Quién lo hizo: Emad Ahmad Abousud, Universidad Islámica Imam Mohammad Ibn Saud, Riad
El límite que importa: Ningún tribunal publicado de esos países ha decidido un caso civil por deepfake; todo es anticipación.

Cómo tratan el daño moral a las empresas dos códigos civiles

Artículo 138(2) de Arabia Sauditafrente aArtículo 267 de Jordania

El texto saudí limita la compensación por daño moral a la persona natural y excluye a las empresas; el texto jordano es más amplio, pero la Corte de Casación lo ata al sufrimiento psicológico de la víctima, lo que también deja fuera a las empresas.

Texto jordano (artículo 267)frente aInterpretación de la Corte de Casación jordana

El texto no restringe a la persona natural, pero la corte lo ancla a la experiencia psicológica subjetiva, y una empresa no tiene psique que sufra.

Nada de esto es ley todavía.
Lectura de weeklyAI

Ver · 0:31

Imagine que alguien usa un programa de inteligencia artificial para hacer un video falso con su cara o su voz. La grabación se sube a una red social, el sistema la muestra a miles de personas y otros usuarios la copian y la vuelven a subir. Usted quiere reclamar por el daño a su reputación.

Un profesor de administración de empresas de Riad, Emad Ahmad Abousud, revisó qué dicen al respecto la Ley de Transacciones Civiles de Arabia Saudita de 2023 y el Código Civil de Jordania de 1976. Para comparar, miró normas de la Unión Europea, Estados Unidos, Emiratos Árabes Unidos y China. Su trabajo se publicó en julio de 2026 en la revista Frontiers in Artificial Intelligence.

En Arabia Saudita encontró un problema de redacción. El artículo 138(2) limita la compensación por daño moral a "la persona natural", es decir, a un ser humano. Una empresa víctima de un fraude con video falso no puede reclamar por el daño a su buen nombre. En Jordania el texto es más amplio, pero los tribunales lo interpretan según el sufrimiento psicológico de la víctima. Una empresa tampoco tiene psicología.

El segundo problema es el número de culpables. Los dos códigos hablan de un solo responsable. Un video falso, en cambio, nace de un modelo que lo genera, alguien que lo pide, una plataforma que lo difunde y desconocidos que lo replican. Además, los datos sobre cómo se generó y cómo se difundió quedan en manos de las plataformas y de los desarrolladores.

El autor propone cambiar el artículo 138(2) para incluir a las empresas, aclarar las reglas de custodia sobre sistemas algorítmicos y trasladar al demandado la carga de probar que no causó el daño.

Nada de esto es ley todavía. El propio autor advierte que ninguna corte publicada de Arabia Saudita o Jordania ha decidido un caso civil por un video falso. Todo lo que describe es una predicción.

Para comparar, revisa normas concretas de la Unión Europea, Estados Unidos, Emiratos Árabes Unidos y China, como la ley europea de inteligencia artificial y una sentencia europea de diciembre de 2025 sobre datos personales.

Queda por ver si los tribunales de esos dos países aceptarán tratar un algoritmo como una cosa bajo custodia, si moverán la prueba al demandado o si leerán el daño moral para cubrir a una empresa.

Qué significa para usted

Si usted es empresa, o representa a una, vale la pena seguir si algún tribunal de Arabia Saudita o Jordania acepta tratar un algoritmo como una cosa bajo custodia de quien lo controla. Si es persona, lo que puede observar es si en esos dos países se permite reclamar por el daño al buen nombre de una compañía. Todavía no hay sentencia publicada, así que conviene esperar.

Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975

Quién pagó: El artículo indica que el trabajo fue apoyado y financiado por la Deanship of Scientific Research de la Imam Mohammad Ibn Saud Islamic University (subvención IMSIU-DDRSP2602); no dice si el financiador tuvo alguna influencia y no menciona equipos ni programas prestados.

Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1429 palabras · unos 7 minLeerla →Cerrar

Un video falso con su cara o el nombre de su empresa: qué falta en dos códigos civiles árabes, y por qué esa falta le importa a cualquiera que viva de su buen nombre

Un análisis de las leyes civiles de Arabia Saudita y Jordania encuentra tres huecos ante los daños causados por videos y audios falsos. No hay un solo caso decidido: todo es anticipación.

Un video donde usted aparece diciendo algo que nunca dijo, o un audio donde su voz pide una transferencia. Ese material ya no es una broma de internet. El artículo que revisamos lo describe como parte de un problema más amplio, el de los contenidos sintéticos, capaz de dañar la dignidad de una persona, la reputación de una empresa, la confianza del mercado y la confianza en las pruebas digitales¹. El autor es Emad Ahmad Abousud, de la Universidad Islámica Imam Mohammad Ibn Saud, en Riad, y el trabajo se publicó en Frontiers in Artificial Intelligence. La financiación provino de la Dirección de Investigación Científica de esa misma universidad, mediante una beca con número de referencia.

Conviene ver primero cómo se fabrica ese material. Los métodos actuales de generación usan redes que aprenden a producir imagen y audio realistas compitiendo entre sí, modelos que manipulan identidad y expresión a partir de fotos, y sistemas que refinan ruido hasta convertirlo en imagen de alta fidelidad; así lo describe el resumen de un estudio sobre detección, del que pudimos leer solo el resumen porque el texto completo está detrás de una suscripción². El mismo resumen advierte que esas técnicas sirven tanto para el entretenimiento como para la manipulación política, la invasión de la privacidad, el robo de identidad, la desinformación y los ataques informáticos³.

El caso que el autor usa como retrato del problema ocurrió en Hong Kong en enero de 2024: una suplantación por video de altos ejecutivos dentro de una empresa⁴. La pérdida fue de más de 25 millones de dólares, pero el daño de fondo que el artículo señala es otro: la confianza de una institución y el valor de su nombre.

¿Qué tan grande es? El informe de fraude de identidad de Sumsub registra un aumento de diez veces en el fraude corporativo facilitado por deepfakes en un solo periodo de doce meses⁵. En el Golfo, la comisión saudí de comunicaciones y analistas independientes de verificación de identidad reportan un alza de 600 por ciento en incidentes de deepfake en el reino solo en el primer trimestre de 2024⁶. Piense en qué significa ese 600 por ciento: no es una cifra de casos resueltos, es una cifra de incidentes que se multiplican en tres meses.

El artículo compara dos códigos civiles, el saudí de 2023 y el jordano de 1976, y encuentra tres huecos. El primero es de quién puede reclamar. El artículo 138(2) saudí limita la indemnización por daño moral a "la persona natural", y esa frase, leída con las reglas de interpretación saudíes, deja fuera a las personas jurídicas, es decir, a las empresas⁷. Lo curioso es que el mismo artículo enumera el honor, la reputación y la posición social, tres cosas que aplican igual a una empresa que a una persona: la norma traza la frontera en el cuerpo y luego la extiende a lo que no tiene cuerpo⁸.

En Jordania el texto es más amplio y no restringe a la persona natural, pero el resultado práctico se parece. La Corte de Casación jordana ha anclado el daño moral a la experiencia psicológica subjetiva de la víctima, y una empresa no tiene psique que sufra⁹. Los dos regímenes, pues, llegan al mismo punto desde textos distintos¹⁰. Y ese punto importa, porque el perfil de litigio dominante es corporativo: el autor sostiene que la restricción no afecta a una categoría marginal de demandante, sino posiblemente a una categoría central¹¹.

El segundo hueco es de cuántos culpables. El video dañino no lo produce una sola mano. Interviene quien entrenó el modelo, quien escribió la instrucción, la plataforma que lo amplifica con su algoritmo y quien lo vuelve a subir cuando ya lo bajaron. Los códigos hablan en singular: "quien cause daño". El autor propone que se aclare que los sistemas algorítmicos de alto riesgo pueden tratarse como cosas que exigen cuidado especial, con base en la custodia efectiva que alguien tiene sobre ellos.

El tercero es de prueba. Si el daño se produjo dentro de un sistema opaco, la víctima no puede ver el registro del algoritmo ni los datos de entrenamiento. El autor propone que, cuando la víctima muestre indicios suficientes, la carga de probar la falta de causalidad pase al demandado. Y deja dicho lo esencial: nada de esto está probado en tribunales. La jurisprudencia publicada de Arabia Saudita y Jordania todavía no parece incluir una sola decisión civil sobre deepfakes, así que el argumento es anticipatorio¹².

Hay un detalle que suele quedar fuera de las noticias: la ley de protección de datos personales saudí no crea por sí misma una acción privada por daño moral; solo ofrece mecanismos administrativos y penales¹³. Es decir, la vía que un particular tiene para cobrar no está ahí.

Así lo leemos nosotros. La identidad de una persona o de un negocio se arma con relatos, y esos relatos producen efectos reales: un cliente que se va, un banco que desconfía, un socio que pide garantías. Cuando la ley solo reconoce el daño a la persona física y no el buen nombre de una empresa, el daño que no cabe en ninguna categoría no desaparece: se queda sin quien lo repare, y quien lo sufre carga con la prueba y con el costo. Esperamos, por eso, que las primeras víctimas corporativas de la región terminen pagando de su bolsillo el hueco legal. Nos equivocaríamos si los tribunales de la zona ya aceptaran demandas de empresas por daño a su buen nombre, o si esas empresas obtuvieran reparación por otra vía sin cambiar la ley.

Y hay un segundo patrón que nos parece útil. Los derechos nuevos no bajan de una lista escrita: aparecen cuando un grupo concreto vive una exclusión y la reclama en público. Esperaríamos que las protecciones frente a la suplantación sintética lleguen primero donde haya víctimas organizadas que documenten el daño, no necesariamente donde la ley esté mejor redactada. Sabríamos que nos equivocamos si la protección aparece solo por decisión de legisladores, sin presión de nadie, o si países con leyes amplias no tienen ningún uso práctico de ellas. De ahí lo que usted puede hacer: guarde pruebas fechadas antes de que circule un video con su cara o el nombre de su empleador, y pregunte por escrito a la autoridad de su país qué vía de reparación existe para una empresa, no solo para una persona.

Para ubicar la discusión: en Brasil, el Tribunal Superior Electoral se reunió en febrero de 2024 para fijar reglas de cara a las elecciones municipales de octubre, y el tema central fue el uso de inteligencia artificial en la publicidad política en audio, video y medios audiovisuales¹⁴¹⁵. Según el resumen de ese trabajo, del que pudimos leer solo el resumen, el tribunal se apoyó en la legislación de la Unión Europea y en las elecciones generales argentinas de 2023¹⁶, y ese proceso argentino se volvió una referencia regional sobre el uso de IA en campañas¹⁷. Es la misma familia de problemas, con otra puerta de entrada: allá se regula por la vía electoral, aquí por la vía civil.

Aquí está lo que este análisis no puede decirle. Es un estudio doctrinal sobre Arabia Saudita y Jordania, y no dice nada sobre lo que permiten las leyes de México, Colombia, Chile, Estados Unidos o Canadá; no puede decirle qué harían sus tribunales. Las reformas que propone son sugerencias del autor, no ley vigente, así que nada de esto es un derecho que usted pueda usar hoy en ningún país. No se midió ningún efecto, ni cuántas víctimas cobraron, ni cuánto bajó el daño. Y el tratamiento de la Unión Europea, Estados Unidos, Emiratos Árabes Unidos y China es selectivo, no un repaso completo de esos sistemas. Lo que sí queda es una pregunta que usted puede llevar a donde viva: si mañana alguien usa su cara para pedir dinero, ¿qué autoridad de mi país recibe esa denuncia, y qué tendría que probar yo para que me reparen el daño al nombre?

Ese video con su cara puede aparecer cualquier día. Lo que usted tenga guardado y anotado ese día, y la autoridad a la que sepa acudir, valdrán más que cualquier explicación posterior.

De dónde sale cada dato de contexto, y cuánto leímos de cada documento

Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "They now form part of a wider problem of synthetic-media harm, capable of damaging personal dignity, corporate reputation, market trust, and public confidence in digital evidence."
Nour TM, Omar IAS, Wasan SA. (2026). Multimedia Deepfake Threats and Detection Strategies Based on Limited Data: An Exhaustive Survey. ITEGAM- Journal of Engineering and Technology for Industrial Applications (ITEGAM-JETIA). 10.5935/jetia.v12i59.3473 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The current state of deepfake generation is informed by recent artificial intelligence methods, such as Generative Adversarial Networks (GANs), which learn to create realistically visual and audio content via adversarial learning; autoencoder-based models, which allow manipulating identity and expression compared to images; and diffusion models, which are trained to produce high-fidelity media by refining noise to data."
Nour TM, Omar IAS, Wasan SA. (2026). Multimedia Deepfake Threats and Detection Strategies Based on Limited Data: An Exhaustive Survey. ITEGAM- Journal of Engineering and Technology for Industrial Applications (ITEGAM-JETIA). 10.5935/jetia.v12i59.3473 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Even though these practices aid in innovative uses in entertainment, availability, and customized media, they also help to perpetrate malpractices like political and social manipulation, privacy invasion, identity theft, misinformation, and cybersecurity assaults."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The January 2024 Hong Kong corporate-fraud incident, involving deepfake impersonation of senior executives ( Hong Kong Police Force, 2024 ), captures the practical shape of the problem."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Sumsub (2024) Identity Fraud Report records a tenfold increase in deepfake-enabled corporate fraud over a single 12-month period."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Within the Gulf, the Saudi Communications, Space and Technology Commission and independent identity-verification analytics report a 600 per cent surge in deepfake incidents in the Kingdom in the first quarter of 2024 alone."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Article 138(2) confines the moral-harm head of damages to “the natural person” ( al-shakhṣ dhū al-ṣifa al-ṭabī’iyya ), and that phrase, read against the established canons of Saudi statutory construction, excludes juridical persons from the protected category."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The remaining three (honour, reputation, social standing) apply with equal force to corporate entities. The provision draws the boundary at corporeality and then extends it outward to capture the non-corporeal heads as well."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The Court of Cassation has anchored the moral-harm provision to the subjective psychological experience of the victim—a construction that effectively excludes corporate plaintiffs from the head, since a corporate entity has no psyche to experience moral suffering."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The two regimes thus converge on outcomes from textually divergent starting points—a convergence that gives empirical force to Pound’s older observation that “law in action” can reverse “law in books.”"
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "It may affect a central category of claimant."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Published Saudi and Jordanian case law does not yet appear to include a direct civil-liability decision on deepfakes. The argument is therefore anticipatory."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - el artículo del que trata esta nota — el artículo completo — el pasaje: "But the PDPL provides administrative and criminal mechanisms; it does not itself create a private cause of action for moral damages independent of the Civil Transactions Law’s general framework."
Pessôa P, Bomfim I. (2026). Regulation of artificial intelligence in digital media to combat disinformation and defend the democratic state. Communication & Society. 10.15581/003.39.1.013 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "In February 2024, the Superior Electoral Court of Brazil met in session to determine rules and regulations to guide the municipal elections of the country in October of 2024."
Pessôa P, Bomfim I. (2026). Regulation of artificial intelligence in digital media to combat disinformation and defend the democratic state. Communication & Society. 10.15581/003.39.1.013 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Among the topics debated, the leading one was the deliberations on illicit acts in electoral advertisements in digital media, concerning the use of artificial intelligence, ensuing in regulations on the use of technology in developing, producing and distribution of political advertisements in digital forms such as: audio, video or audiovisual media."
Pessôa P, Bomfim I. (2026). Regulation of artificial intelligence in digital media to combat disinformation and defend the democratic state. Communication & Society. 10.15581/003.39.1.013 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The article will discuss the deliberations of the Superior Electoral Court, which based its decisions on current legislation in the European Union and in the 2023 general elections in Argentina, regarding production aided by artificial intelligence for electoral campaigns."
Pessôa P, Bomfim I. (2026). Regulation of artificial intelligence in digital media to combat disinformation and defend the democratic state. Communication & Society. 10.15581/003.39.1.013 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Argentina’s general election of 2023 was paramount and now a base of a regional case regarding the use of artificial intelligence will be debated in light of reflections on media and technology from authors such as Santaella (2023), Simon (2022), Chagas (2017) e Bucci (2024)."

Los hallazgos de otros estudios que aquí se mencionan los conocemos por este documento, que fue el que leímos; no abrimos cada uno de esos estudios.

analysis of texts · Frontiers in artificial intelligence · the paper, 2 Jul 2026 · free

A Fake of Your Face, a Fake of Your Company: In Saudi Arabia and Jordan, the Law May Not Help

The article proposes reforms, but no court in either country has yet decided a civil deepfake case.

Short version · the longer version follows, about 7 min

Ask the weeklyAI chat

Ask me about this study: who was studied, what it found, and what it does not say.

Conversations are saved for as long as weeklyAI exists, to improve the publication. Answers come in the language you write in.

The study at a glance
Who: Saudi and Jordanian civil codes, read against EU, US, UAE and Chinese materials
How many: two national civil codes (Saudi Arabia, Jordan)
Where: Saudi Arabia and Jordan
When: Saudi law of 2023; Jordanian code of 1976
Kind of study: analysis of what people did
Who did it: Imam Mohammad Ibn Saud Islamic University, Riyadh
The limit that matters: No published Saudi or Jordanian court has yet decided a civil deepfake case; the argument is anticipatory.

How each code treats a company's claim to moral harm

Saudi Article 138(2), 2023againstJordanian Article 267, 1976

Saudi text names only 'the natural person', shutting companies out; Jordan's text says 'anyone' with no such limit

Jordan's textagainstJordan's courts

The text is broad, but the Court of Cassation ties moral harm to the victim's psychological suffering, which a company cannot have

A right on paper is not a remedy; a remedy is what happens when someone with resources pushes a clause until it moves.
weeklyAI's reading

Watch · 0:52

A video call comes in. On the screen are the faces of people you know: your boss, your colleagues, the person who signs the payments. The faces are fake, produced by a generative model. A platform's recommendation system pushes the video to thousands of viewers. Strangers copy it and upload it again.

For a company that is impersonated, the article says the deeper damage may be to trust and reputation. The author, Emad Ahmad Abousud, a business-administration scholar at Imam Mohammad Ibn Saud Islamic University in Riyadh, read the civil codes of Saudi Arabia and Jordan against this situation.

He finds a gap in the Saudi code. Article 138(2), from 2023, limits compensation for moral harm to "the natural person," so a company cannot claim injury to its reputation under that article. Jordan's Article 267, from 1976, does not name only natural persons. Jordanian courts, though, tie moral harm to a victim's psychological suffering, which a company cannot have. Different words, the same result.

Both codes also picture one wrongdoer. A deepfake involves a model developer, a person who types the prompt, a platform and people who re-upload. The records that would show who did what sit with the platforms and developers.

The article cannot show how any of this plays out. The author says no published Saudi or Jordanian court has yet decided a civil case about deepfakes, so the analysis is a prediction. His proposals are suggestions, not law, and nothing in the article is a right anyone can use today.

The article also looks at how the United States handles these cases. It describes a patchwork of federal and state laws there, and says that model offers little that Saudi Arabia and Jordan could borrow.

What remains unknown is whether courts in Saudi Arabia and Jordan will treat an algorithm as a thing under someone's care, as the author proposes, or shift the burden of proof to a platform, or read moral harm to cover a company. None of it has been tested.

What this means for you

For now there is nothing to file and no right to use in either country: the article proposes changes, and no court there has decided a civil deepfake case. Watch whether a company impersonated by a fake video is allowed to claim harm to its name, and whether a platform is asked to show its records.

Who paid: The article says the work was supported and funded by the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University (grant number IMSIU-DDRSP2602); it does not say whether the funder had any say, and it names no equipment or software lent.

The longer versionThe passages copied from the paper, the pictures, and every source with how much of it we read · 1416 words · about 7 minRead it →Close

# When Your Face Is Faked, the Law May Ask First Whether You Have a Psyche

A legal review of Saudi and Jordanian civil codes finds that a company whose reputation is faked may have no clear route to compensation — and that no court in either country has yet decided such a case.

Consider the January 2024 Hong Kong corporate-fraud incident, in which deepfakes impersonated senior executives and induced fraudulent transfers. The faces were synthetic. The January 2024 corporate-fraud incident, in which deepfakes impersonated senior executives, is the case the article opens on to show what synthetic-media harm looks like in practice¹.

The article, by Emad Ahmad Abousud of Imam Mohammad Ibn Saud Islamic University in Riyadh, is a doctrinal review published in Frontiers in Artificial Intelligence. It reads the Saudi Civil Transactions Law of 2023 and the Jordanian Civil Code of 1976 against each other and against selected EU, US, UAE and Chinese materials. It measured nothing. Its empirical material — incident reports and regulatory documents — is used, in the author's words, in a deliberately modest way. And the author states plainly that no published Saudi or Jordanian case law yet appears to include a direct civil-liability decision on deepfakes. The argument is anticipatory².

What the technology is matters here, because the legal problem follows from how the fakes are made. Recent artificial-intelligence methods generate them: generative adversarial networks, which learn to produce realistic visual and audio content through adversarial learning; autoencoder-based models, which allow manipulating identity and expression; and diffusion models, which refine noise into high-fidelity media — according to the summary of a review of deepfake threats and detection strategies, of which we could read only the abstract, the full paper being behind a subscription³. The same summary notes that these methods aid entertainment and custom media and also help perpetrate political and social manipulation, privacy invasion, identity theft and misinformation⁴.

The first gap the article identifies sits in a single phrase. Article 138(2) of the Saudi code confines compensation for moral harm — the non-money injury to honour, reputation and standing — to "the natural person," and the author reads that phrase, against Saudi rules of construction, as excluding companies and institutions from protection⁵. His reasoning is worth following: three of the heads of harm listed in the provision (suffering, body, freedom) presuppose a body, but the other three (honour, reputation, social standing) apply with equal force to a company. The provision draws its boundary at having a body, then extends that boundary over the heads that have nothing to do with bodies⁶.

Jordan looks better on paper and arrives at a similar place. Its Article 267 speaks of "anyone" whose honour, reputation or financial standing is infringed, with no restriction to natural persons. But the author reports that the Jordanian Court of Cassation has anchored moral harm to the subjective psychological experience of the victim — a construction that effectively excludes corporate plaintiffs, since a company has no psyche to suffer⁷. Two codes, two different texts, one outcome: what the author calls law in action reversing law in books⁸.

The article's second and third gaps concern who can be sued and what can be proved. Deepfake harm passes through at least four layers — the model developer, the person who prompts or uploads, the platform whose recommendation system decides who sees it, and the users who re-upload it. Classical liability doctrine, in both codes, is written in the singular: one wrongdoer, one harm. And the evidence needed to prove how a harm travelled — model architecture, training data, recommendation logs — sits with the defendants.

The author's proposal is a layered model, and it is a proposal, not law. He suggests amending the Saudi provision to cover "a natural or juridical person," adding injury to commercial reputation and what he calls reputational capital⁹. For Jordan he asks not for a new statute but for the Court of Cassation to reverse its own reading, which he argues the text does not compel¹⁰. Around this he adds a custody rule for algorithmic systems, a private right to compensation under the data-protection laws (which he notes currently offer only administrative and criminal mechanisms, not a private claim for moral damages¹¹), cautious shifting of the burden of proof, and labelling duties for synthetic media¹².

Here is how we read it. A right on paper is not a remedy; a remedy is what happens when someone with resources pushes a clause until it moves. The lesson a reader outside these two countries can take is not about Saudi or Jordanian courts. It is about the question to ask of any AI law where you live: who is expected to bring the case, and is there a body — a consumer agency, a union, a press-freedom group, a data regulator — that takes such cases for people who cannot pay a lawyer? You could tell we were wrong if a single person, unrepresented and without an organisation behind them, won a deepfake reputation claim quickly and cheaply. Until you see that, note the names near you before you need them.

We also read the article's four-layer account as a warning about a familiar deflection. When a fake spreads, the company that made the tool and the platform that ranked it can each point at the anonymous uploader, or at the system itself, as if the system were a creature nobody owns. The author's answer is to attach responsibility to control: whoever controls the design, the deployment or the distribution answers for the risk. What you can do with that is small and practical. When you complain about a fake of yourself or your workplace, ask in writing which named company is responsible for the hosting, the ranking and the generation, and keep the answer. A reply that says only "our system did it" is itself worth showing to a regulator or a journalist.

The scale the article cites is not its own measurement, and it should be read as the author's framing of the problem. It points to a tenfold increase in deepfake-enabled corporate fraud over a single 12-month period, citing an industry identity-fraud report¹³, and to a 600 per cent surge in deepfake incidents in Saudi Arabia in the first quarter of 2024 alone, citing the Saudi communications regulator and identity-verification analytics¹⁴. The author's conclusion from that profile is that the corporate victim is not a marginal plaintiff but possibly a central one¹⁵ — and that forcing such a plaintiff to claim only provable money losses systematically under-captures the harm¹⁶.

In February 2024, Brazil's Superior Electoral Court met to set rules for that year's municipal elections, and the leading topic was illicit acts in digital electoral advertising involving artificial intelligence — regulations covering how the technology could be used to develop, produce and distribute political advertising in audio, video and audiovisual form, according to the summary of a study of that regulation, of which we could read only the abstract¹⁷¹⁸. That study reports that the court based its decisions on current European Union legislation and on Argentina's 2023 general elections¹⁹. Useful to know, and not a substitute for a remedy: election rules govern campaigns, not your face in a fraud video.

Three limits deserve stating plainly. The article is about Saudi and Jordanian civil law; it says nothing about what courts in the United States, Canada or any Latin American country would do, so it cannot tell you what your own courts would allow. Its reforms are one scholar's suggestions, not enacted law, so nothing in it is a right you can use anywhere today. And it is a policy and practice review, not empirical research: the comparative treatment of the EU, US, UAE and China is selective, and the case-law search found no directly applicable decisions. The author also warns that reform should not impose open-ended platform liability, suppress lawful expression or discourage legitimate work on generative AI — a trade-off, not a free lunch.

The video call is still the place to end. What the article gives you is not a shield but a checklist for reading the next AI law that reaches your country: does it name a person or company who must answer, does it give you a way to get the evidence held by the other side, and does it let a company claim the harm when its own name is faked? Ask those three of your representatives, and you will know more than the headlines will tell you.

Where each piece of context comes from, and how much of it we read

Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "The January 2024 Hong Kong corporate-fraud incident, involving deepfake impersonation of senior executives ( Hong Kong Police Force, 2024 ), captures the practical shape of the problem."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "Published Saudi and Jordanian case law does not yet appear to include a direct civil-liability decision on deepfakes. The argument is therefore anticipatory."
Nour TM, Omar IAS, Wasan SA. (2026). Multimedia Deepfake Threats and Detection Strategies Based on Limited Data: An Exhaustive Survey. ITEGAM- Journal of Engineering and Technology for Industrial Applications (ITEGAM-JETIA). 10.5935/jetia.v12i59.3473 — only the abstract - the full text could not be fetched — the passage: "The current state of deepfake generation is informed by recent artificial intelligence methods, such as Generative Adversarial Networks (GANs), which learn to create realistically visual and audio content via adversarial learning; autoencoder-based models, which allow manipulating identity and expression compared to images; and diffusion models, which are trained to produce high-fidelity media by refining noise to data."
Nour TM, Omar IAS, Wasan SA. (2026). Multimedia Deepfake Threats and Detection Strategies Based on Limited Data: An Exhaustive Survey. ITEGAM- Journal of Engineering and Technology for Industrial Applications (ITEGAM-JETIA). 10.5935/jetia.v12i59.3473 — only the abstract - the full text could not be fetched — the passage: "Even though these practices aid in innovative uses in entertainment, availability, and customized media, they also help to perpetrate malpractices like political and social manipulation, privacy invasion, identity theft, misinformation, and cybersecurity assaults."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "Article 138(2) confines the moral-harm head of damages to “the natural person” ( al-shakhṣ dhū al-ṣifa al-ṭabī’iyya ), and that phrase, read against the established canons of Saudi statutory construction, excludes juridical persons from the protected category."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "The remaining three (honour, reputation, social standing) apply with equal force to corporate entities. The provision draws the boundary at corporeality and then extends it outward to capture the non-corporeal heads as well."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "The Court of Cassation has anchored the moral-harm provision to the subjective psychological experience of the victim—a construction that effectively excludes corporate plaintiffs from the head, since a corporate entity has no psyche to experience moral suffering."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "The two regimes thus converge on outcomes from textually divergent starting points—a convergence that gives empirical force to Pound’s older observation that “law in action” can reverse “law in books.”"
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "Moral harm includes the physical or psychological suffering inflicted on a natural or juridical person as a result of injury to body, freedom, honour, reputation, or social standing, and includes—in the case of the juridical person—injury to its commercial reputation or its reputational capital."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "The Jordanian counterpart problem requires no statutory amendment but judicial reorientation. Article 267 of the Jordanian Civil Code is textually adequate; the restrictive judicial gloss anchored to subjective psychological experience is not compelled by the text and may be reversed by the Court of Cassation through a guiding decision."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "But the PDPL provides administrative and criminal mechanisms; it does not itself create a private cause of action for moral damages independent of the Civil Transactions Law’s general framework."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "The article’s main contribution is a calibrated model of layered civil liability: clearer protection for juridical-person reputation, a custody-based rule for algorithmic systems, cautious burden shifting where relevant information is controlled by platforms or developers, stronger private-law links to data-protection regimes, and targeted transparency duties for synthetic-media systems."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "Sumsub (2024) Identity Fraud Report records a tenfold increase in deepfake-enabled corporate fraud over a single 12-month period."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "Within the Gulf, the Saudi Communications, Space and Technology Commission and independent identity-verification analytics report a 600 per cent surge in deepfake incidents in the Kingdom in the first quarter of 2024 alone."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "It may affect a central category of claimant."
Abousud, E. A. (2026). Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity. Frontiers in Artificial Intelligence. https://doi.org/10.3389/frai.2026.1873975 - the article this story is about — the whole article — the passage: "Restricting the remedy to provable pecuniary damage forces the corporate plaintiff to litigate in a head of damages that systematically under-captures the harm."
Pessôa P, Bomfim I. (2026). Regulation of artificial intelligence in digital media to combat disinformation and defend the democratic state. Communication & Society. 10.15581/003.39.1.013 — only the abstract - the full text could not be fetched — the passage: "In February 2024, the Superior Electoral Court of Brazil met in session to determine rules and regulations to guide the municipal elections of the country in October of 2024."
Pessôa P, Bomfim I. (2026). Regulation of artificial intelligence in digital media to combat disinformation and defend the democratic state. Communication & Society. 10.15581/003.39.1.013 — only the abstract - the full text could not be fetched — the passage: "Among the topics debated, the leading one was the deliberations on illicit acts in electoral advertisements in digital media, concerning the use of artificial intelligence, ensuing in regulations on the use of technology in developing, producing and distribution of political advertisements in digital forms such as: audio, video or audiovisual media."
Pessôa P, Bomfim I. (2026). Regulation of artificial intelligence in digital media to combat disinformation and defend the democratic state. Communication & Society. 10.15581/003.39.1.013 — only the abstract - the full text could not be fetched — the passage: "The article will discuss the deliberations of the Superior Electoral Court, which based its decisions on current legislation in the European Union and in the 2023 general elections in Argentina, regarding production aided by artificial intelligence for electoral campaigns."

The findings of other studies mentioned here are known to us through this document, which is the one we read; we did not open each of those studies.

The free weekly drip by email

One email a week: the free story of every section, with its picture, and the headlines of the rest.

No payment, nothing to cancel. Unsubscribe from any issue.

El goteo semanal gratis por correo

Un correo por semana: la nota gratuita de cada sección, con su imagen, y los titulares de las demás.

Sin pago y sin nada que cancelar. Puede darse de baja desde cualquier envío.