This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
In April 2018, investigative journalist Rana Ayyub was targeted by a malicious campaign involving an AI-generated deepfake pornographic video. The video was widely circulated on social media, leading to severe harassment, doxxing, and threats against her life. The incident caused significant psychological harm and forced Ayyub to self-censor her professional work, highlighting the real-world dangers of non-consensual deepfake content.
Investigative journalist Rana Ayyub was targeted by a deepfake porn campaign, where AI-generated explicit content falsely depicted her in a pornographic video. This was part of a broader effort to discredit and silence her, which included a doxxing attack that exposed her personal information that resulted in severe harassment and emotional distress.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Also classified as
The incident involved exposing the victim and the broader public to highly toxic, non-consensual pornographic content.
The incident involved the creation of a non-consensual deepfake pornographic video to target, humiliate, and manipulate the reputation of a specific victim.
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 deepfake video was intentionally created and distributed by human actors to target and harass the victim after the AI model had been deployed.
Added categories peak at 4: Severe (Epistemic Harm, Indirect)
A prominent 2018 incident where Indian journalist Rana Ayyub was targeted with an AI-generated deepfake pornographic video, leading to viral harassment, doxxing, and forced self-censorship. While it did not threaten physical infrastructure or state sovereignty, the incident represents a substantial threat to societal stability and human rights, establishing a dangerous precedent for using synthetic media to undermine press freedom.
The AI system was used as a tool by malicious actors to generate a synthetic video based on direct inputs (the victim's face and a source video) without exhibiting autonomous planning or decision-making capabilities.
Risk Level 3: Limited Risk. The incident involves the creation of AI-generated content such as deepfakes, which under the EU AI Act are subject to specific transparency obligations to ensure users are aware of the AI-generated nature of the content.
The AI system was used specifically to insert the victim's face into a pornographic video, which is the definition of deepfake video generation.
The reports provide highly detailed, first-hand accounts from the victim herself, corroborated by secondary analysis of the incident's societal impact. The sequence of events, the nature of the AI's involvement (deepfake generation), and the resulting harms are clearly documented.
None. The incident is unambiguously a targeted harassment campaign using synthetic media.
Technology
Realistic Deepfake Generation
AI models generate highly realistic fake videos that fool human eyes.
Algorithmic Amplification
Platform algorithms prioritize engagement, spreading hoaxes 10x faster.
Lagging Detection Tools
AI detection tools struggle to keep up in the cat-and-mouse arms race.
Data Inputs
Unconsented Facial Images
Public photos of the victim were used to synthesize the deepfake video.
Lack of Content Metadata
Absence of verifiable provenance data on uploaded media files.
Human Factors
Cognitive Bias to Believe Video
Humans naturally believe visual evidence, making deepfakes persuasive.
Malicious Intent of Actors
Political opponents used deepfakes for targeted character assassination.
Tendency to Share Hoaxes
Users are highly inclined to share salacious and negative content.
Process and Methods
Slow Platform Moderation
Social media bans on fakes are not implemented immediately or perfectly.
Ineffective Debunking Processes
Debunking is too slow to catch up with rapidly spreading viral lies.
Management
Inadequate Platform Safeguards
Tech companies failed to proactively block deepfake distribution.
Lack of Accountability
Political organizations failed to stop members sharing deepfakes.
Regulatory Environment
Lack of Police Action
Law enforcement failed to investigate or act against deepfake creators.
Weak Legal Protections
Inadequate legal frameworks to protect women from deepfake abuse.
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.
