Incident 835: La technologie de l'IA alimenterait de fausses informations sur les catastrophes naturelles et les accidents en Chine
Description: En Chine, des outils d'IA auraient été utilisés pour fabriquer et diffuser de fausses informations sur des catastrophes, notamment un glissement de terrain dans le Yunnan, un tremblement de terre dans le Sichuan et une mort subite suite à un accident de la circulation. Le 27 mai 2024, un véritable séisme de magnitude 5,0 s'est produit dans le comté de Muli, au Sichuan, sans faire de victimes et avec des dégâts matériels limités. Cependant, une publication sur les réseaux sociaux a ultérieurement prétendu à tort que l'épicentre se trouvait dans le comté de Xide et a exagéré la gravité de l'événement, en ajoutant des images fabriquées de toutes pièces de destructions massives. Les auteurs de ces fausses informations ont depuis été sanctionnés administrativement par les autorités chinoises.
Editor Notes: Incident 835 presents multiple discrete AI-related incidents, each pointing to the wider application of AI tools to fabricate events and spread misinformation on social media throughout China, similar to Incident 834. Anonymous names like "Moumou" are used to preserve privacy in reporting. The following is a reconstruction of the timeline of events associated with this incident ID: (1) 1/23/2024, a fabricated landslide incident in Yunnan: Yang Moumou allegedly used AI tools to create and share a fake news article claiming that a landslide in Yunnan had resulted in eight deaths. That misinformation was posted on social media to attract views and engagement. Authorities identified the incident as AI-driven misinformation, and Yang was subsequently given an administrative penalty. (2) 5/27/2024, a real 5.0-magnitude earthquake occurred in Muli County, Sichuan. It resulted in minor property damage but no casualties. However, this actual event would later become the basis for an exaggerated AI-fabricated narrative. (3) 6/9/2024, a fabricated earthquake report in Xide County, Sichuan: Luo Moumou reportedly used AI to enhance and circulate a false report about the May 27 earthquake, falsely claiming the epicenter was in Xide County, Sichuan, with severe property damage and casualties. The fabricated images and videos spread quickly on social media, which led to confusion and alarm. Luo received an administrative penalty for disseminating false information. (4) Sometime in June 2024, a fabricated fatal traffic incident in Chengcheng County: Tian Mou used AI to create and post a story claiming that a middle-aged woman in Chengcheng County suffered a fatal heart attack after her electric scooter was impounded during a traffic stop. The narrative was designed to generate an emotional response for the purposes of driving engagement, but it misled the public. Tian admitted to fabricating the story using AI and was issued an administrative penalty.
Entités
Voir toutes les entitésAlleged: Unknown deepfake technology developer et Unknown AI developers developed an AI system deployed by Yang Moumou , Tian Mou et Lou Moumou, which harmed Yunnan general public , Xide County in Sichuan residents , Sichuan general public , Muli County in Sichuan residents , Chinese general public , Chinese citizens et Chengcheng County Shaanxi residents.
Statistiques d'incidents
Risk Subdomain
A further 23 subdomains create an accessible and understandable classification of hazards and harms associated with AI
3.1. False or misleading information
Risk Domain
The Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental harms, and (7) AI system safety, failures & limitations.
- Misinformation
Entity
Which, if any, entity is presented as the main cause of the risk
Human
Timing
The stage in the AI lifecycle at which the risk is presented as occurring
Post-deployment
Intent
Whether the risk is presented as occurring as an expected or unexpected outcome from pursuing a goal
Intentional
Rapports d'incidents
Chronologie du rapport
Ces dernières années, les technologies de l’information représentées par le big data, l’intelligence artificielle (IA), etc. ont évolué chaque jour. Dans le même temps, la technologie de synthèse des outils d'IA est utilisée pour fabriquer …
Variantes
Une "Variante" est un incident de l'IA similaire à un cas connu—il a les mêmes causes, les mêmes dommages et le même système intelligent. Plutôt que de l'énumérer séparément, nous l'incluons sous le premier incident signalé. Contrairement aux autres incidents, les variantes n'ont pas besoin d'avoir été signalées en dehors de la base de données des incidents. En savoir plus sur le document de recherche.
Vous avez vu quelque chose de similaire ?
Incidents similaires
Did our AI mess up? Flag the unrelated incidents
Incidents similaires
Did our AI mess up? Flag the unrelated incidents