Veritensor – open-source tool to scan AI models for malware and license issues (github.com)

Veritensor has launched an open-source security tool specifically designed for analyzing AI models, addressing critical vulnerabilities such as malware and license compliance. Unlike traditional antivirus solutions, Veritensor employs deep abstract syntax tree (AST) analysis and cryptographic verification, enabling it to evaluate AI model formats like Pickle, PyTorch, and Keras. The tool ensures that models are safe from malicious code, authentic against tampering, and compliant with licensing terms, thereby protecting users from risks like remote code execution and man-in-the-middle attacks. This tool is significant for the AI/ML community as it enhances trust and security throughout the AI supply chain, ensuring that only secure and compliant models make it into production environments. Veritensor integrates seamlessly with CI/CD processes, allowing for automated scanning and signing of models. By utilizing features like license firewalls, hash verification against trusted registries, and static analysis for obfuscated attacks, it empowers developers to maintain rigorous security standards while fostering innovation. Its lightweight installation and ease of use make it accessible for various development pipelines, reinforcing the community's commitment to safety and compliance in AI deployment.