Succinct Labs’ ZCAM fingerprints iPhone photos and video

Succinct Labs on Thursday launched ZCAM, an iPhone app that cryptographically signs photos and video at the moment of capture. The app creates a digital fingerprint tied to the specific device so recipients can check whether media came from a real phone and has not been altered or generated by AI.

ZCAM produces a cryptographic hash from the captured pixels using device hardware and embeds a signature that the company describes as unique and tamper-proof. Succinct described the app as “signing photos and videos at the moment of capture, producing a tamper-proof record that links content to the device that captured it.” Recipients can independently verify those signatures without relying on a central authority.

Succinct presented the device-based method as an alternative to commercial AI-detection tools. The company reported internal tests indicating some AI detectors can fail to identify generated or altered content and said linking signatures to hardware provides a different verification signal.

The company highlighted potential uses for businesses and news organizations that need to establish provenance and authenticity for images and footage. Succinct also acknowledged that the system requires people and organizations to capture and sign media with ZCAM for the signatures to be useful in practice.

ZCAM is built on Succinct’s existing cryptographic infrastructure, including an SP1 zero-knowledge virtual machine the company says currently secures more than $4 billion in digital assets. Succinct launched the mainnet for its Succinct Prover Network last August, a decentralized marketplace on Ethereum where applications submit zero-knowledge proof requests and independent provers compete to validate them.

The company activated a native PROVE token and raised $55 million in 2024 in a financing round led by Paradigm, with participation from the founders of Polygon and EigenLayer. Succinct said the Prover Network and zkVM handle proof tasks that support the signatures and verification steps, and that multiple independent provers can increase resilience in the validation process.

Succinct cited research projecting that generative AI could raise U.S. fraud losses from $12.3 billion in 2023 to $40 billion by 2027. The company noted other projects are also exploring blockchain-based identity and human verification systems intended to help distinguish real people and real media from AI-generated content.

Adoption challenges remain. The signatures are effective only when content is captured and signed with ZCAM, and the company did not announce partnerships with platforms or newsrooms at launch. Integrating provenance checks into editorial and business workflows and creating verification habits among platforms and audiences are additional hurdles.

The company positioned ZCAM as a provenance tool rather than a content removal mechanism. By creating a verifiable record at capture, the app aims to provide a way to confirm device origin and capture time for photos and video that carry its cryptographic signatures.