Michael S. Levine and Alex D. Pappas | Hunton Andrews Kurth
It is no secret that insurance stakeholders are waiting with bated breath to see how the insurance industry will address the risks and opportunities posed by artificial intelligence (AI). As stakeholders consider various insurance products and their coverage for AI risks, the definition of AI is vital. The reason is that if the bounds of AI are not delineated, then efforts to either grant or restrict insurance coverage for AI are likely to be futile.
As influential Enlightenment thinker John Locke explained long ago: “So difficult it is to show the various meanings and imperfections of words when we have nothing else but words to do it with.” This statement has proved particularly prescient in the insurance field. Because insurance policies must, by their nature, reduce abstract and often complex concepts to writing, the various meanings and imperfections inherent in certain words have been, and will always be, a vexing issue for insurers and policyholders alike.
Examples of disputes about the meaning of specific words in the insurance field are wide-ranging. For example, does the word “automobile” include an 18-wheeler? Are jetskis “watercrafts?” Does property sustain “physical loss or damage” when a harmful substance or virus changes the condition of the property such that it no longer functions as intended? Does the phrase “because of” mean any causal connection at all or is it limited to foreseeable causal relationships? Are there circumstances where the word “or” can actually mean “and” in context? The examples of wordchoice are many, but the point is the same: without a stable, predictable and limited definition of a given term or phrase, disputes about the scope of insurance coverage are all but certain to arise with real-world consequences.
AI presents the latest frontier in this centuries-old definitional quest. On the one hand, the definitional problem is the same as with other words in that there is inherent difficulty in distilling abstract intellectual concepts in writing. On the other hand, AI poses new challenges due to its technological complexity and novelty. That is, there are at least seven forms of AI that presently exist or could exist and it is essential that insurance policy definitions adequately capture which is included and excluded. Compound this complexity with the reality that the average consumer of commercial insurance will have little understanding (if any) as to which forms of AI are being utilized in the insured business operations:
1. Reactive Machines AI: These are the simplest AI systems. They are purely reactive and can neither form memories nor use past experiences to inform current decisions.They are meant to perform specific tasks and their behavior is deterministic.
2. Limited Memory AI: These AI systems can learn from historical data to make decisions. They can store past experiences or data for a brief time. An example of this is self-driving cars that observe other cars’ speed and direction.
3. Theory of Mind AI: This is a more advanced type of AI that can understand thoughts and emotions that affect human behavior. They can interact socially. This type of AI currently exists only in theory.
4. Self-Aware AI: This is the final stage of AI development which currently exists hypothetically. Self-aware AI, which currently exists only in theory and science fiction, would be systems that have their own consciousness and self-awareness.
5. Artificial Narrow Intelligence: Also known as Weak AI, this type of AI is meant to perform a narrow task, such as voice recognition. These systems can only learn or be taught how to do specific tasks.
6. Artificial General Intelligence: Also known as Strong AI, this type of AI refers to a system that possesses the ability to perform any intellectual task that a human can do. They can understand, learn, adapt and implement knowledge in a broad range of tasks.
7. Artificial Superintelligence: This refers to a time when the capability of computers will surpass humans. ASI is currently a hypothetical concept often depicted in science fiction. It is proposed to be capable of extraordinary cognitive capabilities, including the ability to understand and master any intellectual task that a human can do.
Considering these AI variants, the question for insurance industry participants is how to achieve contractual certainty while maintaining reasonable scope. The answer to this question can have wide-ranging, multi-billion-dollar implications.
Apart from the inclusion or exclusion of certain categories of AI, the definition of AI ultimately used in insurance policies will have to grapple with technologies or business processes that only use AI as a single subcomponent. In other words, insurance policy terms will have to be drafted to account for AI-enabled or AI-adjacent technologies—not technologies that solely use AI.
In sum, it remains to be seen how insurers will define AI in new policy forms. But as definitions are formulated and proposed, it is important to consider all definitional angles so that all involved parties can have as much certainty as possible about insurance products being bought or sold.
When one of your cases is in need of a construction expert, estimates, insurance appraisal or umpire services in defect or insurance disputes – please call Advise & Consult, Inc. at 888.684.8305, or email experts@adviseandconsult.net.