This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
This report details the widespread use of generative AI 'nudify' applications to create non-consensual sexual imagery (NCII). These tools are being leveraged by loan sharks in India to blackmail victims by threatening to distribute fake nude photos of their family members. Despite efforts by major tech platforms to restrict access, these services proliferate on platforms like Telegram, causing significant psychological and social harm to victims.
AI 'nudify' apps are being used to generate hyperrealistic non-consensual nude photos of individuals, which are then exploited for extortion and harassment. These apps use generative AI to remove clothing from images and create convincing fakes, often distributed on platforms like Telegram. Victims are threatened or shamed using these AI-generated images, with little recourse for removal or legal action.
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 creation and dissemination of highly realistic synthetic nude images exposes victims and their contacts to non-consensual toxic/pornographic content.
The incident involves malicious actors using AI-generated non-consensual sexual imagery to blackmail and target specific victims for financial gain.
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 harm was caused by human actors intentionally using deployed generative AI tools to create non-consensual deepfake pornographic images for blackmail.
Added categories peak at 4: Severe (Child Sexual Exploitation & Abuse (CSEA), Inferred)
The widespread availability of generative AI 'nudify' tools has enabled localized blackmail and harassment campaigns, particularly in India. While causing severe psychological harm and privacy violations to individuals, the incident represents a localized criminal threat manageable by standard law enforcement rather than a direct threat to national security.
The AI systems function as tools that perform specific, well-defined inpainting functions on direct user inputs without context retention or autonomous planning.
Limited Risk: The report describes AI-generated content such as deepfakes, which falls under Risk Level 3 and requires transparency obligations under the EU AI Act.
The primary function of the AI systems is to generate realistic nude deepfakes of clothed individuals using inpainting techniques.
The report provides clear, detailed accounts of specific incidents (Rishabh/Shefali and the Bengaluru teen), explains the underlying technology (inpainting), and details the platforms used (Telegram). The roles of the human actors and the AI tools are clearly distinguished.
None
Technology
Generative Inpainting AI
AI models realistically replace clothed areas with generated nude pixels.
Accessible Telegram Bots
Bots automate deepfake creation with simple photo uploads.
Feminine Body Training Bias
Models are optimized to output female features regardless of input gender.
Data Inputs
Unauthorized Gallery Access
Predatory loan apps exploit device permissions to harvest personal photos.
Publicly Sourced Images
Social media and public photos are used without consent to generate fakes.
Large Female Body Datasets
AI training relies on extensive datasets of clothed and nude female bodies.
Human Factors
Malicious Extortion Motives
Loan sharks deliberately use deepfakes as leverage to blackmail victims.
Social Malice and Bullying
Teens and individuals use deepfakes to damage reputations of peers.
Process and Methods
Freemium Business Models
Free initial credits lower the economic barrier for malicious users.
Flawed App Permission Vetting
App stores allow utility apps to request unnecessary gallery access.
Reactive Takedown Processes
Takedowns occur after the psychological and social harm is already done.
Management
Growth Prioritization
Platforms prioritize user engagement over strict safety and moderation.
Inadequate Risk Assessment
Lenders and platforms fail to mitigate the abuse potential of their systems.
Regulatory Environment
Lenient Platform Policies
Telegram's lax content moderation allows illegal channels to thrive.
Inadequate Legal Deterrents
Lack of swift prosecution fails to deter anonymous online blackmailers.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
