Key facts
- AIUC has developed a third-party audit and certification layer for AI agents.
- The company's AIUC-1 standard is inspired by the cybersecurity standard SOC 2.
- AIUC tests AI agents for issues including jailbreaks, hallucinations, and data leaks.
- Customers include Cursor, Lovable, Harvey, and ElevenLabs.
- AIUC uses AI agents to run tests, with humans verifying the final audit.
Artificial Intelligence Underwriting Company (AIUC), founded by former Anthropic employee Rune Kvist and ex-METR COO Rajiv Dattani, has secured $40 million in Series A funding to provide third-party audits for AI agents. The startup aims to address enterprise concerns about the safety and reliability of AI systems, particularly as they become more sophisticated.
The company's approach is modeled after cybersecurity standards, with its AIUC-1 standard and testing service designed to validate AI agents against risks such as jailbreaks, hallucinations, and data leaks. AIUC assembles a consortium of security and risk leaders to shape its testing protocols, which involve running agents through approximately 5,000 tests. The results are compiled into a detailed report, with final verification conducted by humans, though AI agents are used for the testing and analysis phases.
AIUC counts Cursor, Lovable, Harvey, and ElevenLabs among its customers. The $40 million Series A round was led by Ribbit Capital, with participation from First Harmonic. This follows a $15 million seed round, bringing AIUC's total funding to $55 million. The company's efforts align with broader industry calls for increased AI safety and pacing of development, as highlighted by Anthropic CEO Dario Amodei.