Markets with Teeth
SF 00:00NYC 00:00LDN 00:00

AIUC builds standards, audits and insurance for frontier AI.

Every big technology wave needs trust infrastructure. Electricity burnt down many houses before Underwriters Laboratory discovered how to make lightbulbs safe. Today, every lightbulb in America carries UL’s trustmark. Similarly, Visa created trust in electronic money, Moody’s created trust in railroad bonds, and FICO built trust algorithmic lending.

We help businesses and governments answer: “Can we trust this agent?” As agents become smarter than humans, trust becomes the bottleneck. Trust needs independent perspectives.

Artificial Intelligence Underwriting Company creates trust infrastructure: standards, audits and insurance. Standards define controls that create trust, audits stress-test that the controls work, and insurance puts money on the line. AIUC-1, the technical standard for agent security, is built with security leaders at the world’s biggest banks, hospitals and technology company. Frontier agent builders use AIUC-1 to earn trust.

Rune Kvist & Rajiv DattaniCo-founders

A technical partner to the world’s fastest growing companies and most esteemed global institutions.

Standards don’t enforce themselves. Markets create teeth.

  1. Standards

    Define the requirements. Specify safeguards such as data access controls, prompt injection defenses, and tool call authentication. Explore AIUC-1

  2. Audits

    Put agents to the test. Attempt jailbreaks, probe for data leaks, and check whether safeguards hold up under pressure. View an Audit Report

  3. Insurance

    Put money on the line. Insurers bear the cost when AI fails, giving them a financial stake in standards and audits that reduce real risks. View the ElevenLabs Policy

StandardsDirectsAuditsInformsInsuranceUpdatesMarkets with Teeth

AIUC-1: The technical standard for agent security, safety, and reliability

Built with 300+ security & risk leaders, AIUC-1 addresses the risks slowing AI adoption. Independent audits verify that agents meet its requirements, giving businesses confidence to put AI to work.

The trustmark for AI agentsBuilt with 300+ security & risk leaders
  1. Data & Privacy

    Protect data and information we entrust to AI. Limit access to sensitive data and guard against leaks of customer information and trade secrets.

  2. Reliability

    Check that AI can actually do the work we expect. Test for hallucinations, verify tool calls, and build safeguards against errors that can proliferate throughout the business.

  3. Security

    Defend against manipulation of AI systems. Test resistance to prompt injections, restrict permissions, and prevent agents from taking actions beyond their authorized scope.

  4. Accountability

    Make ownership and responsibility explicit. Assign owners, maintain records of AI activity, and establish response plans so people can investigate failures and intervene when needed.

  5. Safety

    Reduce the risk of harm. Test for dangerous and inappropriate outputs, enforce boundaries on agent behavior, and route high-risk situations to human review.

  6. Society

    Address risks that extend beyond any one business. Establish safeguards against AI enabling cyberattacks and other dangerous capabilities, with attention to the potential for harm at scale.