ProvCam: A Camera Module with Self-Contained TCB for Producing Verifiable Videos

Our perception of reality is under constant threat from ever-improving video manipulation techniques, including deepfakes and generative AI. Therefore, proving authenticity of videos is increasingly important, especially in legal and news contexts. However, it is very challenging to prove it based on post-factum video content analysis. In this work, we take a preventative stance and construct ProvCam, a novel camera module that generates a cryptographic proof of video authenticity. Our solution greatly reduces the size of Trusted Computing Base (TCB) to include the module itself. Moreover, it mitigates tampering during the numerous processing steps between video capture by the camera sensor and generation of the digital video output. To confirm its practicality, we present a complete prototype of ProvCam on a Xilinx FPGA evaluation board. As experiments show, ProvCam incurs a negligible performance overhead (latency and throughput) and small energy consumption overhead when recording a video. It imposes a moderate hardware cost but is relatively small compared to other major components such as SoC. Moreover, it does not change the existing camera software stack and thus can be easily integrated with various camera-bearing devices, such as smartphones.