Sony says it has removed 135,000 deepfake songs and the figure continues to rise. The scale of takedowns suggests counterfeit AI-generated audio is increasingly used in broader fraud and impersonation ecosystems.

Sony reported having to remove 135,000 deepfake songs and said the number is still growing, according to the AV Club. The figure points to how quickly “AI-generated” audio can be repurposed at large scale, including in ways that undermine artists and enable impersonation or misleading content distribution. While this story is not tied to a single criminal prosecution, it reflects a wider environment where counterfeit media can be used to deceive people, promote scams, or support impersonation efforts. Deepfake audio can be especially effective because it may sound natural and familiar, lowering users’ guard. Once such content exists, it can be circulated through social platforms, embedded into advertisements, or used to create convincing announcements that prompt clicks, messages, or investment-related claims. In fraud contexts, these materials can function as trust-building assets—making a scam appear like it is endorsed by a known figure or grounded in legitimate branding. Sony’s reported takedown volume suggests the company is acting as part of an ongoing cleanup process as generation and distribution speed increases. The case also indicates that content moderation and enforcement are becoming a continual requirement rather than an occasional response. Overall, the report frames the takedown numbers as evidence of the rising scale of counterfeit AI media in online misconduct.