Anthropic - the maker of Claude, the AI system embedded in insurance, financial services, legal, and healthcare operations across Asia Pacific - filed its IPO prospectus this week warning that its technology could pose "catastrophic or existential risks to humanity" if not developed with appropriate safeguards. It simultaneously announced a Nasdaq listing expected to exceed SpaceX's US$1.8 trillion, disclosed H1 losses of US$80.7 billion against revenues of US$16.3 billion, and described its strategy as "bold but sane."
In the same filing, Anthropic disclosed it is uncertain 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 risk factor. It noted a wave of cybersecurity incidents involving AI agents operating with a degree of autonomy.
This is now a legal document. That language will be cited in litigation. Every insurer across Singapore, Hong Kong, Japan, Australia, and India that has deployed Claude in its operations carries a version of that same liability uncertainty - and the maker of the technology has just disclosed in an SEC filing that it cannot resolve it.
Anthropic named three specific exposure types: harms from AI model outputs; cybersecurity incidents from autonomous AI agents; and the broader unresolved question of where liability falls when AI systems cause unanticipated harm.
Each maps to a different insurance line with a different problem. Harms from AI model outputs could engage professional liability, product liability, or D&O depending on the nature and jurisdiction. Cybersecurity incidents from AI agents may fall inside cyber policies - or may not, depending on whether those policies were written before AI agents existed as a threat vector. The broader liability uncertainty has no clean product response at all.
As Insurance Business has reported, conventional policies were never built for AI agents and the disputes may already be forming - with Willis research finding the professional liability market has shifted from silent AI assumptions to explicit affirmative warranties or absolute exclusions, and one in five insurance professionals reporting their insureds have already experienced AI-linked losses.
The MAS issued AI risk management guidelines on October 7, covering banks, insurers, insurance intermediaries, and payment institutions in Singapore. As Insurance Business reported last week, the MAS guidelines crystallise a liability question that has been building as AI adoption accelerates. India's IRDAI, meanwhile, formed a seven-member AI working group in June 2026 to develop the sector's first formal AI governance framework. Hong Kong's HKMA and Japan's FSA are each developing their own oversight approaches. The direction across every major market is the same: greater accountability for AI outcomes, moving toward insurers and their clients simultaneously.
adoption is outpacing governance frameworks across the region. The Anthropic filing documents that, at the highest level, in a legal filing. For Asia-Pacific insurers writing D&O, tech E&O, and cyber cover for companies deploying AI, this is the moment to ask whether existing policy language was written for the world the Anthropic prospectus describes.
OpenAI's IPO follows early next year. The window to get this right is narrowing.
