This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
AI-generated deepfake images of singer [name withheld] attending the 2024 Met Gala circulated on social media, causing confusion among fans and family members. The subject publicly clarified that the images were fake and that she did not attend the event. The incident highlights the potential for synthetic media to deceive individuals and the public.
Purported AI-generated images depicting Katy Perry attending the 2024 Met Gala circulated widely on social media, leading to public confusion about her presence at the event. The images were realistic enough to deceive fans and even Perry’s mother, who believed them to be authentic. Perry publicly clarified the images were fake, attributing them to AI and noting she had missed the event due to work.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
The incident involved the generation and viral spread of a highly realistic deepfake image of a celebrity, which misled the public and the celebrity's family into believing she attended an event she did not.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by an unknown human actor intentionally generating and distributing deepfake images of a celebrity using a deployed AI image generator.
Added categories peak at 4: Severe (Epistemic Harm, Inferred)
AI-generated deepfake images of singer Katy Perry at the 2024 Met Gala circulated widely on social media. While demonstrating the increasing realism of synthetic media, the incident was benign in nature, targeting a celebrity for entertainment purposes, and carries negligible national security implications.
The AI system functioned as an image generation tool, producing a static synthetic image based on user inputs without exhibiting autonomous decision-making or planning capabilities.
Risk Level 3. Limited Risk: The system generated a deepfake, which is subject to specific transparency obligations under the Act to ensure users are aware they are interacting with or viewing AI-generated content.
The AI system was used to generate a highly realistic, synthetic image of a real person (Katy Perry) to depict her at an event she did not attend.
The report clearly details the incident, the parties involved, and the nature of the AI-generated content. There is no ambiguity about the fact that the image was a deepfake and that it deceived the subject's mother.
None
Technology
Realistic Image Generation
AI generators can create highly realistic fake images of public figures.
Lack of Provenance Markers
Generated images lacked metadata or watermarks identifying them as AI-made.
Data Inputs
Abundant Celebrity Training Data
Abundant public images of Katy Perry allowed for highly accurate training.
Human Factors
Cognitive Bias and Credibility
Public expectations of her attending made the fake highly believable.
Low Digital Literacy
Users and family members failed to recognize visual indicators of AI synthesis.
Process and Methods
Inadequate Platform Labeling
Social media lacked automated detection and labeling for AI-generated media.
Rapid Viral Dissemination
Sharing mechanisms allowed the image to spread before verification occurred.
Management
Prioritizing Virality Over Safety
Platforms prioritized user engagement over implementing verification checks.
Regulatory Environment
Absence of Deepfake Laws
Lack of legal frameworks governing non-consensual synthetic celebrity media.
AI system involved
An AI system was involved, but the reports do not identify it.
Identified by automated extraction; can differ from the alleged developers above.
