As artificial intelligence becomes more deeply embedded in business decision-making, questions of accountability, liability, and insurability are moving from theoretical concern to practical urgency. For risk managers, insurers, and corporate leaders alike, the challenge is no longer whether AI introduces new risks, but how those risks are understood, governed, and transferred in an environment where legal, regulatory, and technical frameworks are still evolving.
This conversation with Professor Anat Lior, conducted as part of Willis Towers Watson’s ongoing research into AI liability through the Willis Research Network, builds on earlier analysis in Insuring the AI Age and offers a timely deep dive into how liability for AI-related harm is beginning to take shape. Drawing on empirical legal research and close engagement with insurers, regulators, and technology developers, the discussion highlights where existing insurance models are being stretched, where new approaches are emerging, and where uncertainty remains.
The themes below are particularly relevant for organisations deploying or underwriting AI systems today, as they speak directly to coverage gaps, regulatory divergence, litigation risk, and the evolving role of insurance in enabling safe and responsible AI adoption.
This recording dates to Summer 2025. Viewers should be aware that certain reports, policies, or legislative frameworks discussed may have been updated since that time.
01
AI risk is not yet fully categorized within traditional insurance frameworks. The empirical research conducted by Professor Lior reveals a field in flux: some insurers and brokers remain cautious, citing a lack of claims data and relying on existing tech or cyber policies, while others, especially startups and innovative departments, are proactively developing AI-focused solutions. For risk managers, this means the market is still experimenting, and coverage gaps or ambiguities may exist, particularly for novel AI applications outside established domains like autonomous vehicles.
02
The regulatory environment for AI is rapidly evolving but remains fragmented. The EU AI Act is poised to reshape compliance expectations, but its practical enforcement and impact on insurance are still unclear. In the US, the absence of unified regulation creates further uncertainty. Risk managers can monitor both legislative developments and insurer responses, as regulatory shifts could quickly alter liability exposures and policy requirements.
03
Traditional actuarial models may not fully capture the risks associated with AI, particularly as new technologies and use cases emerge. The conversation highlights that while some risks can be approached with existing methodologies, others, such as those arising from generative AI or agentic systems, require new thinking. Guarantee policies, which focus on performance failure rather than accident-based liability, are emerging as one response. Risk managers can assess whether their current policies address the unique characteristics of AI risk or if bespoke solutions are warranted.
04
The legal landscape is shifting, with high-profile cases (e.g., involving generative AI and copyright, or AI-related harm) influencing underwriting and policy design. Insurers are increasingly attentive to litigation outcomes, which may set precedents for coverage and compensation. For risk managers, staying informed of litigation trends is essential for anticipating potential claims scenarios and ensuring adequate coverage.
05
Insurance can play a key role in enabling safe AI adoption by providing a financial safety net for unforeseen outcomes. However, the conversation also highlights the need for greater collaboration between insurers, technology experts, and regulators to ensure that insurance products keep pace with technological change. Risk managers may advocate for clear policy language regarding “silent AI” risks and seek affirmative statements from insurers about AI coverage. Engaging in industry forums and cross-sector dialogues can help organizations stay ahead of emerging risks and regulatory expectations.
06
Emerging technologies such as quantum computing are expected to further complicate the risk landscape. The convergence of AI and quantum will introduce new uncertainties, requiring risk managers to remain agile and informed. The future may see the development of standalone AI policies or the absorption of AI risk into broader insurance products, depending on market and regulatory evolution.
Taken together, these insights underline a central message for risk managers and insurers: AI liability is not a distant or abstract concern, but a present and dynamic risk that requires active engagement. While the insurance market is beginning to respond through experimentation, endorsements, and emerging product concepts, much of the landscape remains unsettled, particularly for novel and rapidly evolving AI applications.
WTW’s collaboration with Professor Lior reflects a broader commitment to helping clients navigate this uncertainty with clarity and foresight. By combining legal scholarship, market insight, and practical risk advisory expertise, this work aims to support organisations in anticipating liability exposure, engaging constructively with insurers, and shaping governance frameworks that enable innovation without undermining accountability.
As AI technologies continue to advance and intersect with other emerging risks, including quantum computing, the ability to align legal understanding, insurance solutions, and risk management practice will be critical. Ongoing dialogue across industry, academia, and policy will play a key role in ensuring that insurance remains a meaningful enabler of safe, resilient, and responsible AI deployment.
