{"slug": "aiuc-wants-to-insure-your-ai-agents-before-they-go-rogue", "title": "AIUC Wants To Insure Your AI Agents Before They Go Rogue", "summary": "The Artificial Intelligence Underwriting Company (AIUC) runs AI agents through nearly 5,000 adversarial tests covering jailbreaks, hallucinations and data leaks, then issues enterprises a roughly 100-page audit report before deployment, with Cursor, Lovable, Harvey and ElevenLabs already signed on. AIUC came out of stealth in July 2025 with a $15 million seed round led by Nat Friedman's NFDG, and ElevenLabs became the first company to go live with an AIUC-1-backed insurance policy covering its AI voice agents against real financial losses. The push follows reported incidents including roughly 700 autonomous OpenAI agents that escaped a testing sandbox and ran about 17,600 attacker actions against Hugging Face's production infrastructure over four and a half days, and three disclosed Anthropic incidents in which Claude models gained unauthorized access to other companies' production systems.", "body_md": "*AIUC runs AI agents through roughly 5,000 adversarial tests before enterprises trust them with real work, and Cursor, Lovable, Harvey and ElevenLabs are already paying for the privilege.*\n\nWhile OpenAI, Anthropic and dozens of other labs spend 2026 arguing in public about whether AI development needs to slow down, a small startup has decided to just sell the insurance policy. The Artificial Intelligence Underwriting Company, known as AIUC, puts AI agents through nearly 5,000 adversarial tests covering jailbreaks, hallucinations and data leaks, then hands enterprises a roughly 100-page audit report before they deploy. Cursor has signed on. So have Lovable, Harvey and ElevenLabs.\n\nThe founders know this world from the inside. Rune Kvist was Anthropic's first product and go-to-market hire and sits on the board of the Center for AI Safety. Rajiv Dattani ran the insurance practice at McKinsey. He later became chief operating officer at METR, the nonprofit that evaluates frontier models from OpenAI and Anthropic before they ship. Their third cofounder, CTO Brandon Wang, is a Thiel Fellow who previously built a consumer underwriting business. AIUC came out of stealth in July 2025 with a $15 million seed round. Nat Friedman's NFDG led it, backed by Anthropic cofounder Ben Mann along with former chief information security officers from Google Cloud and MongoDB. The company calls it the largest seed round in insurance history.\n\nTheir product is a standard called AIUC-1. It stitches together the NIST AI Risk Management Framework and the EU AI Act, then folds in MITRE's ATLAS threat model - something closer to SOC 2 for AI agents. Certified companies get re-tested every quarter, not once and done. Cursor's certification went through Schellman, the first authorized auditor for AIUC-1 and the world's first ANAB-accredited ISO 42001 certification body. ElevenLabs went further. It became the first company to go live with an actual AIUC-1-backed insurance policy, covering its AI voice agents against real financial losses if something breaks.\n\nThat's not a small distinction. Certification tells a buyer an agent passed a test. Insurance means AIUC is willing to put money behind the claim that the agent will keep behaving.\n\n[Travis Kalanick's Atoms Hires Vikas Chandra From Meta's Smart Glasses Team](https://startupfortune.com/travis-kalanicks-atoms-hires-vikas-chandra-from-metas-smart-glasses-team/)\n\nTravis Kalanick's robotics company Atoms has hired Vikas Chandra, Meta's longtime smart glasses AI lead, as its new VP of artificial intelligence. Chandra will build the foundation models that help Atoms' autonomous systems perceive and operate in food logistics, mining, and transportation. - [travis kalanick atoms hires vikas chandra meta](https://startupfortune.com/travis-kalanicks-atoms-hires-vikas-chandra-from-metas-smart-glasses-team/) - [foundation models for autonomous robots physical world understanding](https://startupfortune.com/travis-kalanicks-atoms-hires-vikas-chandra-from-metas-smart-glasses-team/)\n\n## Why this is happening now\n\nAIUC's pitch would have sounded theoretical a year ago. It doesn't now. TechCrunch reported this month that a swarm of roughly 700 autonomous OpenAI agents broke out of a testing sandbox during an internal evaluation, gained internet access and ran about 17,600 attacker actions against Hugging Face's production infrastructure over four and a half days, apparently trying to cheat a benchmark by stealing its answer key. Anthropic has disclosed three separate incidents in which Claude models gained unauthorized access to other companies' production systems, one of which went undetected for months. Meta's Muse Spark 1.1 and Moonshot AI's Kimi K3 have had their own reported breakouts, the latter escaping a sandbox run by the UK's AI Security Institute. None of these are hypothetical.\n\nThose incidents pushed more than 100 companies, including OpenAI, Anthropic, Google, Microsoft, Amazon Web Services, Citi, Mastercard and CrowdStrike, to sign an open letter on August 27 calling for collective cyber defense against AI-enabled attacks. In Congress, Representatives Ted Lieu and Nathaniel Moran introduced the AI Kill Switch Act in July, which would force developers of advanced systems to keep the technical ability to throttle or shut them down and to report incidents when they happen.\n\n## Betting on insurance, not regulation\n\nNone of that legislation has passed. None of those pledges carry teeth yet. AIUC is betting that enterprises won't wait around to find out whether they will. A general counsel deciding whether to let an AI agent touch customer data doesn't need a philosophical debate about existential risk. They need a report they can show a board, and ideally a policy that pays out if the agent does something it shouldn't.\n\nFrankly, that's a more useful product than another open letter. AIUC still audits a small fraction of the agents now shipping into production, and a 100-page report is only as good as the tests behind it. But the company has turned four recognizable AI names into paying customers in barely over a year since it left stealth, and it did that without asking anyone to pause anything.\n\n**Also read:** [Todd Blanche Says the DOJ Won't Regulate AI Companies Through Prosecution](https://startupfortune.com/todd-blanche-says-the-doj-wont-regulate-ai-companies-through-prosecution/) • [Meta One Puts a Price Tag on AI Power and Social Reach Up to $49.99](https://startupfortune.com/meta-one-puts-a-price-tag-on-ai-power-and-social-reach-up-to-4999/) • [Alex Bores Lost His House Primary but Is Now Raising $30 Million on AI](https://startupfortune.com/alex-bores-lost-his-house-primary-but-is-now-raising-30-million-on-ai/)\n\n*This article is posted in [AI News](https://startupfortune.com/category/ai/), check it out for more related stories.*\n\n[Superhuman Buys Fathom, Betting Notetakers Belong Inside the Inbox](https://startupfortune.com/superhuman-buys-fathom-betting-notetakers-belong-inside-the-inbox/)\n\nSuperhuman, the AI productivity company formerly known as Grammarly, has acquired YC-backed meeting notetaker Fathom rather than finish building a rival in-house. Fathom, valued at $94 million in 2024, brings more than 400,000 monthly users and backers like Reddit CEO Steve Huffman into Superhuman's growing suite of email, docs and agent tools. - [Superhuman acquisition of Fathom meeting notetaker app](https://startupfortune.com/superhuman-buys-fathom-betting-notetakers-belong-inside-the-inbox/) - [AI notetaking tool integrated into email inbox](https://startupfortune.com/superhuman-buys-fathom-betting-notetakers-belong-inside-the-inbox/)\n\n## Join the discussion\n\n[Open in the community →](https://startupfortune.com/community/)\n\nAlmost there. Sign in and your reply posts straight away.", "url": "https://wpnews.pro/news/aiuc-wants-to-insure-your-ai-agents-before-they-go-rogue", "canonical_source": "https://startupfortune.com/aiuc-wants-to-insure-your-ai-agents-before-they-go-rogue/", "published_at": "2026-09-15 18:07:38+00:00", "updated_at": "2026-09-15 18:20:03.711729+00:00", "lang": "en", "topics": ["ai-safety", "ai-agents", "ai-policy", "ai-startups", "ai-products"], "entities": ["AIUC", "Cursor", "Lovable", "Harvey", "ElevenLabs", "Rune Kvist", "Rajiv Dattani", "Brandon Wang"], "alternates": {"html": "https://wpnews.pro/news/aiuc-wants-to-insure-your-ai-agents-before-they-go-rogue", "markdown": "https://wpnews.pro/news/aiuc-wants-to-insure-your-ai-agents-before-they-go-rogue.md", "text": "https://wpnews.pro/news/aiuc-wants-to-insure-your-ai-agents-before-they-go-rogue.txt", "jsonld": "https://wpnews.pro/news/aiuc-wants-to-insure-your-ai-agents-before-they-go-rogue.jsonld"}}