Entrepreneurs from anthropic and metr launch AI safety startup with $55m in funding

Rune Kvist and Rajiv Dattani, an early Anthropic employee and former METR COO, have launched Artificial Intelligence Underwriting Company (AIUC), a startup focused on AI safety for enterprises. The company has secured $55 million in funding, including a $40 million Series A led by Ribbit Capital and a $15 million seed round from Nat Friedman’s NFDG. AIUC aims to apply cybersecurity standards to AI risks by developing a certification framework called AIUC-1, inspired by SOC 2. The startup conducts extensive testing of AI agents through 5,000 scenarios, using AI to run tests and analyze data, with human verification for final audits. Dattani emphasized the need for transparency, stating, 'Here’s where it passes and where you can trust it.' AIUC’s approach mirrors METR’s work, which previously focused on performance testing for frontier labs. Anthropic CEO Dario Amodei has also called for third-party safety evaluations. AIUC’s model seeks to provide enterprises with independent assessments of AI agent safety, addressing concerns about rogue behavior and data leaks. The startup’s clients include Cursor, Lovable, Harvey, and ElevenLabs, reflecting its growing influence in the AI safety space.