Proof of Capture: Apple Reference Image, but open source and using steganography (merybenavente.me)

A recent innovation in image authenticity has emerged with the creation of an open-source project called Proof of Capture, developed by Alex Hornstein and a collaborator during their time at the Recurse Center. This technology employs steganography to embed an invisible, signed perceptual hash within the image pixels, allowing photographers to cryptographically prove that their photos were genuinely captured by a camera rather than generated by AI. By shifting the focus from detecting fake images to verifying real ones at the moment of capture, this approach aims to strengthen trust in visual content amid the growing concerns about AI-generated imagery. The Proof of Capture system utilizes an ATECC608 cryptographic chip to securely sign the images, ensuring the private key remains within the chip, inaccessible even to the owner. Unlike Apple's similar initiative that stores signatures in a "digital negative" outside the image itself, which relies on Apple's Private Cloud Compute for authentication, Proof of Capture emphasizes transparency and public accessibility. While neither this system nor Apple's solution, nor existing standards like C2PA, completely eliminate the risk of forgery (such as capturing screens displaying AI images), the open-source nature of Proof of Capture allows hobbyists and professionals alike to build their own devices for under $100, fostering a community-driven approach to reaffirming the integrity of photos.