Early Anthropic hire, former METR COO have found a way to rein in rogue AI agents
techcrunch ·
Rune Kvist and Rajiv Dattani, founders of the startup Artificial Intelligence Underwriting Company (AIUC), have developed a solution to help enterprises manage the risks associated with rogue AI agents. The company recently announced a $40 million Series A funding round led by Ribbit Capital, bringing its total capital raised to $55 million. AIUC aims to bring safety standards to the AI industry by creating a third-party audit and certification layer for AI models and agents. Drawing inspiration from the cybersecurity standard SOC 2, they introduced a new framework called AIUC-1 along with a comprehensive testing service. This service subjects AI agents to approximately 5,000 tests involving jailbreaks, hallucinations, and data leaks to generate detailed safety reports. Although AI uses AI to conduct these tests and analyze data, human experts verify the final audit results to ensure accuracy. The founders leverage their backgrounds as an early Anthropic employee and former METR COO to address growing concerns about AI control in corporate environments.
techcrunch