A major upgrade is a new ability for the detection tool to recognize synthetic faces generated by advanced AI models, such as advanced diffusion and GAN technologies, in addition to traditional deepfakes such as identity and expression swaps. The model claims to offer 97 percent improvement in error rates compared to version 1.0. The tech also provides scores for specific types of manipulation, for more in-depth analysis of the prevalence and nature of deepfake attack vectors. The new deepfake detection toolkit is available through Paravision’s cloud-ready SDKs and Docker-based architecture, making it easy to deploy for digital onboarding, secure access, travel identity, and remote verification environments. Paravision recently launched Liveness 2.0, an update to its liveness detection solution, which among 15 selected to represent the state of the art by the U.S. Department of Homeland Security’s Science and Technology Directorate (S&T).