According to the Financial Times in the UK, due to insurance companies becoming more cautious about emerging artificial intelligence (AI) related risks and refusing to provide full coverage, OpenAI and Anthropic are considering using investor funds to address potential litigation costs that could reach billions of dollars.
Sources said that the insurance coverage OpenAI has obtained for "emerging AI risks" is only about $300 million, far less than what is needed to deal with litigation costs of billions of dollars. To this end, the company is evaluating a self-insurance or establishing a "self-insurance" entity plan, intending to draw funds from its nearly $60 billion investor fund pool to cover potential compensation and legal expenses.
Meanwhile, Anthropic has started using internal resources—preparing for a settlement of about $1.5 billion in copyright class-action lawsuits initiated by authors. Previously, a federal court in California had preliminarily approved this copyright settlement involving a large amount of pirated text, and if not resolved, the estimated damages were once extremely high.
Industry insiders said that insurance companies are increasingly wary of so-called "nuclear verdicts"—unprecedented huge compensations that may occur for young technology companies—considering these systemic and unforeseen risks exceed the coverage of traditional policies. The insurance gap may force major AI companies to bear more legal risks with their own capital or investor funds, while also changing the landscape of industry risk management and investor due diligence.
As of the time of writing, OpenAI, Anthropic, and their insurance brokers (such as Aon) have not yet made public comments on the details mentioned above.
