Insights

DeepStrike Estimates Online Deepfakes Surged From 500,000 in 2023 to 8 Million as Voice Cloning Crosses Threshold

Siwei Lyu, a computer science professor and director of the Media Forensic Lab at the University at Buffalo, warned that 2026 will be the year real-time reactive deepfakes fool people at scale. He noted that voice cloning has officially crossed the "indistinguishable threshold" for average users. According to cybersecurity firm DeepStrike, online deepfakes have skyrocketed from 500,000 in 2023 to 8 million in 2025. This rapid growth of nearly 900% annually is accompanied by an increase in retail fraud, where major businesses are receiving over 1,000 AI scam calls per day. Lyu explained that temporal consistency models like OpenAI's Sora 2 and Google's Veo 3 have eliminated traditional tells. He emphasized that the line of defense must shift to cryptographic signing, as looking closely at pixels is no longer sufficient.