Ensuring Authenticity Instead of Detecting Manipulation

Methods that were once considered reliable - such as metadata analysis or visual plausibility checks - are no longer enough to prevent fraud at scale. Even AI-based forensics provides only statistical indications, not legally robust proof. The question is: what can be done?

With ImageSign, Tallence introduces a different approach: instead of proving manipulation, the goal is to make authenticity verifiable. In other words, content authentication replaces heuristic detection. By applying cryptographic principles, both the origin and integrity of an image can be secured. ImageSign is built around three core elements:

• Digital signatures and hashing: An image is uniquely signed at the moment of capture or initial import. Any subsequent modification invalidates the signature.
• Provenance data (C2PA): Origin, editing history, and device integrity are documented in a tamper-resistant manner.
• Trusted Execution Environment (TEE) & device attestation: Ensure that signatures are generated within a secure and trustworthy environment.

This makes it possible to verify whether an image can be authentic - not just whether it appears plausible.

At the same time, it will not always be feasible to rely exclusively on signed images. For third-party, unsigned content, insurers must continue to apply multi-layered forensic analysis on a case-by-case basis. ImageSign is designed to support these scenarios as well, incorporating:

• JPEG error level analysis (ELA) to detect localized manipulations
• Neural inconsistency detection (e.g., CNNs, vision transformers) for anomalies in lighting, shadows, and materials
• Metadata forensics (EXIF data, toolchains, camera profiles)
• Copy-move analysis to identify duplicated regions
• Physical plausibility checks (light sources, perspective, shadows)
• Detection of retouching based on edge and gradient anomalies

Against this backdrop, the only truly robust approach to deepfake detection is to secure images cryptographically at the moment of creation. With ImageSign, a verifiably unaltered piece of evidence is created - one that cannot be compromised, even by highly convincing fakes.

The key is that authenticity is not validated retrospectively in backend systems but is technically guaranteed during the capture process itself. Ultimately, the combination of secure image capture, a hardened device environment, and protected metadata prevents images from being altered without detection.

To ensure a secure, end-to-end proof of image origin, ImageSign prioritizes the following features:

• Signature at capture using Secure Enclave / Trusted Execution Environment (TEE)
• Immutable binding of timestamp, location, device, and capture parameters
• Protection against device tampering (root/jailbreak detection)
• Storage in secure, C5-compliant environments
• API-based integration into claims and fraud management systems
• Full audit trails and real-time verification

The combination of multiple technical components provides the foundation for a secure, cross-platform, and future-proof implementation that meets high standards of security and integrity. ImageSign delivers these capabilities within a single platform:

• Cross-platform compatibility: The application runs seamlessly on both iOS and Android. Platform-independent technology is essential for broad adoption and consistent security.
• Digital image signatures: ImageSign is capable of digitally signing images using a private key. This signature ensures both the integrity and authenticity of the images.
• Secure key storage: The private key is stored in a secure environment that prevents unauthorized access or extraction by external applications or mechanisms.
• Protection against reverse engineering: This ensures that neither the algorithm nor the cryptographic keys can be extracted through analysis of the application.
• Secure application access: A robust authentication mechanism is integrated, including passwords, biometrics, or other security-relevant methods, allowing controlled access to the application.
• Encryption in transit: Ensuring secure transmission of images and digital signatures is critical. ImageSign achieves this through the implementation of encryption protocols, particularly in communication with servers or external services.