Anthropic, the developer of Claude, has brought these questions into sharper focus through disclosures in its initial public offering (IPO) prospectus. The document warns that its technology could pose “catastrophic or existential risks to humanity” if developed without appropriate safeguards. It also acknowledges uncertainty over where and how the company might be held legally responsible for actions taken by its systems.
The disclosures have implications beyond the technology sector. Insurers, banks, financial services providers, legal firms and healthcare organisations across the region are incorporating AI into their operations. If these systems produce harmful outputs, expose sensitive information or act in unexpected ways, the resulting losses could trigger disputes over insurance coverage and responsibility.
Anthropic’s prospectus identifies regulatory action arising from AI-related harm, including suicide and child safety concerns, as a potential risk. It also highlights cyber security incidents involving AI agents capable of operating with a degree of autonomy.
Such disclosures could become relevant in future litigation, particularly where claimants seek to establish whether a technology developer, a business using the system or another party should be held accountable. However, the prospectus does not settle those questions. Instead, it underlines the legal uncertainty facing companies that develop and deploy increasingly capable AI systems.
Three categories of risk create distinct insurance challenges
The risks identified in the filing raise three broad questions for insurers: how to cover harm caused by AI-generated outputs, how to respond to cyber incidents involving autonomous agents, and how to allocate liability when an AI system causes an unforeseen loss.
Each presents a different challenge for existing insurance products.
AI-generated outputs can cause financial or reputational damage, particularly when businesses rely on automated analysis or recommendations in professional services. Depending on the circumstances, claims could involve professional indemnity, product liability or directors’ and officers’ (D&O) insurance. The relevant policy would depend on the nature of the alleged wrongdoing, the parties involved and the jurisdiction in which the claim is brought.
Cyber insurance presents a separate set of questions. AI agents can perform tasks with limited human supervision, potentially increasing the number of ways in which an organisation’s systems and information may be exposed to risk. If an agent contributes to a data breach, unauthorised action or other security incident, the policyholder may need to establish whether the event falls within the wording of its cyber policy.
Coverage cannot be assumed. Policies written before autonomous AI agents became a prominent source of concern may not address their activities explicitly. Insurers and policyholders may therefore disagree over whether a particular incident is covered, especially where exclusions or definitions do not clearly account for AI-related events.
The third challenge is the allocation of liability when an AI system behaves in an unexpected manner. A single incident could involve questions about the technology developer’s design, the deploying organisation’s safeguards and the decisions made by people overseeing its use. Existing insurance products may not provide a straightforward solution to every possible combination of these responsibilities.
Insurance policies face pressure to adapt
The insurance market is already responding to concerns about AI-related losses. Research by Willis indicates that the professional liability market has moved away from leaving AI exposure largely unaddressed towards the use of explicit affirmative warranties or complete exclusions.
The distinction matters. An affirmative provision can expressly establish the circumstances in which certain AI-related risks are covered, while an exclusion can remove specified exposures from protection. Both approaches make the contractual position clearer, although the practical outcome depends on the precise wording.
The same research found that one in five insurance professionals reported that their insured clients had already experienced AI-linked losses. This suggests that AI-related claims are not merely a theoretical concern, even though their nature and the extent of coverage disputes will vary between businesses and jurisdictions.
For companies using AI, reviewing insurance arrangements should involve more than checking whether a policy mentions artificial intelligence. They also need to understand how the policy treats data breaches, professional errors, third-party claims, technology failures and losses arising from automated decisions.
Insurers, meanwhile, must consider whether their underwriting assumptions reflect the way AI systems are actually being deployed. Clear definitions, carefully drafted exclusions and explicit coverage provisions can help reduce uncertainty, although they cannot eliminate every legal dispute.
Regulators across Asia-Pacific tighten their focus
Regulatory developments across major markets indicate a growing emphasis on AI risk management and accountability.
Singapore’s Monetary Authority issued AI risk management guidelines on 7 October, covering banks, insurers, insurance intermediaries and payment institutions. The guidance reflects the need for financial institutions to manage the risks associated with adopting AI while maintaining appropriate oversight of their systems and operations.
In India, the Insurance Regulatory and Development Authority of India established a seven-member AI working group in June 2026 to develop the insurance sector’s first formal AI governance framework.
Hong Kong’s Monetary Authority and Japan’s Financial Services Agency are also developing their approaches to AI oversight. Although the details differ between jurisdictions, the broader direction is towards stronger accountability for the outcomes produced by AI systems.
For insurers, these developments create a two-sided challenge. They must manage the risks associated with their own use of AI while assessing the exposures of corporate clients that deploy the technology. A company using AI to process claims, analyse financial information or support professional decisions may face questions about its governance arrangements if the system produces harmful or unreliable results.
Regulatory expectations can also influence underwriting decisions. As oversight becomes more developed, insurers may need to examine whether clients have adequate controls, human supervision and procedures for responding to AI-related incidents.
A narrowing window for insurers to review coverage
AI adoption across Asia-Pacific is advancing faster than the development of consistent governance and liability frameworks. Anthropic’s disclosures illustrate the scale of the challenge: even the companies building advanced AI systems acknowledge uncertainty about the legal consequences of their technology’s actions.
For insurers providing D&O, technology errors and omissions, professional indemnity and cyber cover, the immediate question is whether existing policy language adequately addresses these emerging exposures.
A comprehensive review would examine how policies treat AI-generated harm, autonomous agents, data security incidents and claims involving multiple parties. Insurers would also need to consider whether their clients have appropriate safeguards in place and whether responsibilities between technology providers and users are clearly defined.
The prospect of further major AI company listings adds urgency to the debate. OpenAI’s IPO is anticipated early next year, according to the source material, although the timing remains a forward-looking expectation rather than a confirmed event.
The central issue for the insurance industry is not simply how quickly AI develops, but whether liability rules and insurance contracts can keep pace. As AI becomes more deeply embedded in essential business operations, uncertainty over responsibility and coverage could become an increasingly significant source of disputes, financial exposure and regulatory scrutiny.
