Artificial Intelligence (AI) is a general purpose technology with applications in practically every industry and is expected to contribute trillions of dollars to global GDP in the next 10 years. However, AI also ranks high on the list of risks facing economies, societies, and businesses globally.
So, we may have a bit of a bind here: for the true economic potential of AI to be realized, there needs to be much wider adoption; but for much wider adoption, AI risks must be mitigated effectively. One of the time-tested levers of mitigation that we have is insurance. Can AI insurance break this impasse?
For example, when we consider the large-scale adoption of internet and digital technologies, some credit is due to the insurance industry for providing risk cover against the losses of cyber threats and attacks. Similarly, can the insurance industry underwrite the risks that may accompany AI products and services?
Need of insurance for AI
AI insurance can certainly help if an organization is concerned about liability and thus wary of adopting AI technologies. But before that, why does AI need special coverage? Don’t existing insurance products for cyber insurance, technology errors & omissions insurance, or general liability cover AI risks as well? Well, to an extent yes. But these risk covers have been created prior to the rapid advance of AI capabilities. So, as AI risk incidents increase (for example, if AI-based cyber scams proliferate) insurers may exclude that scenario or make it an add-on endorsement.
Also, as AI is deployed in high-risk use cases, that may warrant additional coverage for those selling AI software. Generative AI brings forth additional types of risk such as intellectual property rights concerns. Some AI software vendors are currently indemnifying their clients against copyright claims and in future, they may look for specialists to reinsure such risks. In short, AI developers/vendors, and businesses using AI software, including in-house developed AI technologies, all can benefit from covering their AI risks.
AI Insurance Opportunity
In fact, as AI gains mainstream adoption, we are seeing the emergence of new insurance products as well. Savvy-insurers who can accurately underwrite AI risks can earn billions in premiums. Deloitte estimates that this category will grow at 80% annually to reach 4.8 billion dollars by 2032. Reinsurance major Munich Re has been offering AI risk coverage since 2018, which now also covers generative AI products.
As noted above, the potential clients include technology vendors, product startups, organizations leveraging AI, etc.
Challenges of underwriting AI Risks
While this is an exciting opportunity, there are several challenges for the market to develop fully. Actuaries need historical data and estimates for successful underwriting. We don’t yet have data about AI risks, probability of incidents, cost of incidents and so on.
Next, insurers can underwrite risk because they have gained a high degree of expertise in the areas they are covering. But for many areas of AI, today that expertise lies with AI companies themselves. The fast-changing nature of the field increases the difficulty of keeping ahead in the field.
However, the ability of evaluate AI systems can only go up as organizations gain more experience. Some regulations require independent audits and third-party assessments of high-risk AI systems. As such audit methodologies become standardized, insurers can also leverage them. In the long term, AI risks may become part of standard coverage or for some specific high-risk scenarios may be excluded or require separate endorsements.
Conclusion
Beyond the growth opportunity for insurers, AI insurance serves another purpose. It reassures the public about AI safety. When experts and specialists with skin in the game have assessed the AI systems and insure against the damages arising from their usage, it’s a signal that they see them as fit for purpose and then the general public can have greater confidence in the safety of such AI systems. In that sense, AI insurance can be a public good!
References
1) Providing insurance coverage for artificial intelligence may be a blue ocean opportunity, Deloitte Financial Services.
2) Insuring Generative AI: Risks and Mitigation Strategies (pdf), Munich Re Whitepaper.
3) Insurance for AI: Easier Said than Done, John Loeber, founder, Limit.
4) Investigating an insurance-for-AI startup, L Rudolf L and Florence Hinder.
{Kashyap Kompella is the CEO of the global technology industry analyst firm RPA2AI Research and a best-selling author. You can check out his books: Practical Artificial Intelligence: An Enterprise Playbook and A Short & Happy Guide to Artificial Intelligence for Lawyers}
Comments