Anthropic - the maker of Claude, the AI system embedded in financial services, insurance, legal, and healthcare operations across the UK - filed its IPO prospectus this week disclosing that its technology could pose "catastrophic or existential risks to humanity" if not developed with appropriate safeguards. It simultaneously announced plans to list on the Nasdaq at a valuation expected to exceed SpaceX's $1.8 trillion summer listing. The company disclosed losses of more than $80.7 billion in the first half of 2026 against revenues of $16.3 billion, and described its strategy as "bold but sane."
The same week, owners of Elvis Presley's music filed suit against Anthropic over the allegedly unauthorised use of his songs to train Claude - a reminder that the IP liability questions around AI training data are already in active litigation, not future speculation.
Anthropic admitted uncertainty about "where and how it might be held liable for actions taken by its systems." It named regulatory action arising from AI model harms - including suicide and child safety - as a material risk factor. It noted a wave of cybersecurity incidents involving AI agents operating with a degree of autonomy.
The coverage gap for AI system liability has been discussed in theory for years. It is now formalised in a legal document by the most prominent AI safety company in the world - a company founded specifically on the premise that AI is dangerous enough to require dedicated safety research. When Anthropic's agents cause harm, that document will be exhibit A in determining who knew what, and when.
As Insurance Business reported in June, PwC found the wage premium for AI skills in UK financial services stands at 53%. The firms building AI accountability structures are capturing that premium. Those that are not are accumulating unpriced liability.
The Anthropic filing lands in the week the Bank of England and FCA published their joint survey confirming 86% of UK financial services firms now use AI, with the median insurer running 18 AI systems in production. The FCA is positioning itself as an "agentic regulator" for an increasingly autonomous financial system. As Insurance Business reported previously, regulators and AI developers are moving toward the same reckoning from opposite sides. The insurance market is in the middle – and that creates a pricing problem.
OpenAI's IPO follows early next year. The window to get policy language right before those claims arrive is narrowing.
