Prof. Siwei Lyu on the Future of Deepfakes
Siwei Lyu, SUNY Distinguished Professor and Empire Innovation Professor of Computer Science and Engineering, featured in an article explaining that deepfakes have advanced dramatically.
University at Buffalo researcher Siwei Lyu, SUNY Distinguished Professor and Empire Innovation Professor of Computer Science and Engineering, warns that deepfakes have advanced dramatically and will become even more convincing in 2026. In a recent article, he explains that AI can now create highly realistic faces, voices and full-body performances that are often indistinguishable from real people, even to trained observers. The number of deepfakes online has also surged into the millions, increasing risks of fraud, misinformation and scams.
Lyu predicts that the next wave of deepfakes will be real-time, interactive synthetic people who can respond during live video calls. Because simply spotting visual flaws is no longer enough, he argues that stronger system-level protections — such as cryptographic media verification and advanced forensic detection tools — will be essential to protect the public from deception.
