Insuring Generative AI: Risks and Mitigation Strategies

Could insurance be the unexpected key to safer AI?

Balancing creativity and responsibility to enable adoption.

This approach could offer several key benefits:

Enhanced Risk Research: Insurance companies with a financial stake in AI would be motivated to conduct thorough research into AI risks and potential harms. This research would be essential for accurately pricing their policies.

Incentivizing Safety: By offering lower premiums to companies that implement safer AI practices, insurers could create a powerful financial incentive for responsible AI development.

Fostering a Safer AI Ecosystem: Insurance could become a market-driven solution to promote AI safety, aligning economic interests with ethical AI advancement.

Could insurance be a key tool in creating a safer AI landscape? What are your thoughts on using insurance as a strategy for managing AI risks?

To this day, AI risks are seldom excluded in traditional insurance policies. Therefore, the damage that AI models cause could be covered by traditional insurance policies. Examples that come to mind are AI-based machinery injuring bystanders (could be covered by existing general liability policies), AI models that are hacked (could be covered by existing cyber insurance policies), AI-based cleaning robots that destroy property (could be covered by existing property insurance policies), AI models that make biased employment decisions (could be covered by existing EPLI policies), and many more.

Insuring Generative AI: Risks and Mitigation Strategies

Due to AI impacting all “walks of life”, there is a lot of partial coverage from existing insurance policies, making it difficult for both insurer and insured to have full confidence on the extent of the coverage – potentially leading to over- or underinsurance. Furthermore, the lack of conclusiveness on the extent of the coverage could lead to devastating effects for insurers, as the threat of “silent AI” might be underrated, as well as for insureds, as they might be left without financial protection. AI exposures within traditional insurance policies could represent a significant unexpected risk to insurers’ portfolios, as the risks of AI underperforming – potentially even systematically – were not considered in the pricing of the insurance.

When using AI, insureds should therefore be aware of potential insurance gaps, leaving them exposed to risks caused by their AI models. Insurers, on the other hand, should be aware of the risks that AI poses to their existing policies and should monitor the risk of silent AI exposure.

Outlook

As once famously said by Henry Ford when talking about New York City: “Without insurance we would have no skyscrapers, because no man would dare to work at such heights, at the risk of killing himself and leaving his family destitute. Without insurance, no businessman would invest his millions in constructing a building like this, when a single spark could reduce it to ashes”. We are still at the beginning of fully encompassing all AI risks, their dependence, their systematic nature and their geographical spread. Similar to the quote above, the risks of GenAI as well as their systematic nature could prove devastating for AI providers or AI users who come to face a class action suit in the US with many damaged parties. Insurance could be the right vehicle to mitigate those risks. Pooling the risks of GenAI going wrong enables active innovation and growth by reducing the risk costs for single GenAI applications and allowing companies to focus on further pushing the technological barriers in GenAI without worrying about the residual financial risks. Furthermore, through the risk assessment function of insurance, much-needed industry standards could be developed, making the AI environment safer without unnecessary overregulation. This creates further trust in the market, allowing us to fully harness the power of AI and GenAI for society and businesses.

The emergence of GenAI has undoubtedly revolutionised industries, offering transformative opportunities while introducing novel risks. With clear guardrails set into place for the use cases, continuous performance monitoring and clear metrics for the model to be measured against, insuring GenAI will be possible as it is insuring prediction performance of ML models. Some of the risks are more easily insurable (and quantifiable) than others.

With the rise of these new and often complex technology risks, a broad demand for insuring GenAI will arise. Some insurance companies are ready for the demand and are looking forward to exploring the reality of insuring GenAI with corporate, broker and primary insurance partners.

Author

AiUTOMATING PEOPLE, ABN ASIA was founded by people with deep roots in academia, with work experience in the US, Holland, Hungary, Japan, South Korea, Singapore, and Vietnam. ABN Asia is where academia and technology meet opportunity. With our cutting-edge solutions and competent software development services, we're helping businesses level up and take on the global scene. Our commitment: Faster. Better. More reliable. In most cases: Cheaper as well.

Feel free to reach out to us whenever you require IT services, digital consulting, off-the-shelf software solutions, or if you'd like to send us requests for proposals (RFPs). You can contact us at [email protected]. We're ready to assist you with all your technology needs.

© ABN ASIA