AI-generated deepfake doctors on TikTok were discovered spreading false medical advice to millions of viewers, with one video receiving 3 million views and 36,000 comments requesting purchase information for unproven health remedies.
Multiple AI-generated avatars posing as medical doctors were discovered spreading false health advice on TikTok, using phrases like '13 years as a coochie doctor' to establish credibility. These deepfake personas, created using an app called Captions, claimed expertise in various medical fields including gynecology, plastic surgery, and nutrition. One particularly viral video featuring an AI avatar named 'Violet' promoted pineapple cucumber salads for gut health and compared lemon balm to Ozempic, accumulating nearly 3 million views. The video generated over 36,000 comments from users requesting recipes and purchase information for the promoted remedies. TikTok user Javon Ford exposed the deception by demonstrating how the Captions app allows users to select AI avatars and input scripts to create convincing medical advice videos. The same AI avatars appeared across multiple accounts claiming different medical specialties, with some also selling wellness products or sharing celebrity gossip. Experts warn these deepfakes are becoming increasingly difficult to detect, with telltale signs including unnatural mouth movements, audio lag, oversized eyes, and awkward speech patterns. The Captions app reportedly has 100,000 daily users producing over 3 million videos monthly.
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
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