A.I. and Insurance; Business Considerations for the Future

Over the course of the last 12 months Artificial Intelligence tools have crossed into mainstream awareness.

No matter what you may think of A.I., as a tool it is poised to significantly disrupt a multitude of industries. From Healthcare and Finance, to Journalism, Creative Writing, Music, and Art, the emergence and ease of access to relatively sophisticated pseudo intelligence programs will have a fundamental impact on the entire world. This development comes hot on the heels of discussions regarding the suitability of human generated content and social media protections, and has some American legislators concerned to the point that they are considering enacting laws to fully regulate Artificial Intelligences moving forward.

As with any new tool, there are a host of considerations and risks for businesses in relation to using an A.I. With the development of any new technology there will be legal and regulatory challenges. It will take time for any given technology to become fully established under law, and during this period there will be significant risk to businesses and professionals using A.I. tools.

Artificial Intelligence and Insurance

Insurance, historically, has created products which in turn have normalized advancing technologies. From the early days of insuring marine cargo policies in London, through to the creation of Fire, Health, and Liability insurance, these products generally arose out of need for their existence.

Prior to their creation, no one had any of these forms of protection.

At one point in history a business owner would have gone to the coffee house and asked the men who sold insurance “My workers injure themselves and I’m always having to pay them compensation above their salary, how can you help me?” and consequently, Employees’ Compensation Insurance would have been born.

While we’re in a slightly more sophisticated age, a similar process is occurring in front of our eyes with regard to Artificial Intelligence – there is a degree of risk in the use of these tools which may not be fully understood at present. But, when a loss stemming from an A.I. platform does, inevitably occur, a market will be born.

At present it should be noted that no standard products exist for covering Artificial Intelligence insurance risks, any A.I. insurance policy in force at the time of publishing this article is likely to be:

1. Bespoke: made specifically for the company exposed to the risk.
2. Highly defined: covering very specific risks related to the tool.
3. Expensive: the fallout of the risk is relatively unknown.

Although we don’t know exactly what losses and risks exist (or will come into existence) because of A.I. we can make some reasonable presumptions about some likely outcomes and suitable products. But there are many questions that will likely to be answered through years of legal challenges first.

Artificial Intelligence: Errors and Omissions?

This week, the website BuzzFeed.com announced that it would begin to use A.I. to generate content. The human writers of the company did not react favorably to this news.

BuzzFeed is in the business of generating content. Producing pop-culture video essays and listicles for a very specific demographic. The content is not critically sensitive. If a mistake is made it is likely not going to be the end of the world. But, an A.I. learns what it wants and is not inherently trustworthy.

So, the question stands: If and when an A.I. makes a mistake that causes a loss, who is responsible?

When dealing with human workers, any advice giving business (including media companies) should have a Professional Indemnity Insurance policy. This ensures that the workers (and business) are protected against the risk of giving bad advice/information, or making a mistake. Can an A.I. be insured in the same way?

Who is to say that the A.I. is qualified or competent to provide the advice it is giving? And even if an A.I. can pass the bar exam, with a C+ average any law firm engaging a modern A.I. is ensuring a mistake probably is going to occur sooner rather than later. 

So, when it inevitably happens, who is to blame for the A.I.’s mistake? 

The company using the A.I., the individual or business which programmed it, or the A.I. itself? To which entity does the party experiencing a loss go to for satisfaction and compensation? Which one holds or is covered by the Errors and Omissions Insurance plan?

An argument could be made that all 3 of these entities are at fault, or maybe even all 3 together. Maybe the blame could be placed at the feet of the A.I. as an “independent entity.”

Can Artificial Intelligence be a “worker?”

If, hypothetically, an A.I. is responsible for its content then does it count as a “worker?”

We’ll ignore the issue of a Turing test and passing for human, but in a future where an A.I. has made a mistake and is being held responsible for its mistake by the organization using it (maybe not as crazy a thought as we might like to think), would it not then follow that the A.I. counts as a “worker?”

That brings up a whole host of issues like – should the A.I. be compensated (maybe with new ram or an operating system) if it breaks due to over work? This may seem outlandish, but there will come a time when a specific class of insurance will likely need to be created to cover the risks of an A.I. “worker.”

There has already been significant discussion regarding the insurance and regulation of artificial intelligence workers, even as early as 1996. Under the “Turing Registry” idea, only entities registered with the system would be eligible for coverage – programmers who want to develop an A.I. would need to receive some form of standardized certification, secure the relevant insurance coverage, and pay a premium before deploying it.

However, this system would rely on how we define an A.I., which then brings us back to the issue of whether we can class an artificial intelligence as a “worker.” 

Even if we could, establishing AI platforms as “workers” means that businesses and employers (possibly even an ordinary consumer) would have to think about the AI as a worker, and would need to treat it as such. This includes possibly even covering the AI with worker’s compensation – it really depends how this issue is going to be defined moving forward.

The Future of Insurance with Artificial Intelligence

At present it seems like the entire world is in a “wait and see” position when it comes to Artificial Intelligence. While businesses worldwide are using A.I. it is still very much an unknown quantity.

For example, insurers are using A.I. to underwrite products, manage claims, and service customers. We can expect the insurance industry worldwide to look very different by the end of the decade simply through the tools that are being created and used. A.I. will change how people live and work, but the machine is not infallible – mistakes will be made and losses will be realized, no matter the quality of the tool you’re using. The technology simply hasn’t advanced to omniscience – yet.

So, any business using an Artificial Intelligence tool should be aware of how that product impacts their risk. Without insurance that is designed specifically to address the work of an A.I. businesses could be left self-insuring their technology; which can be cripplingly expensive should a worst case scenario realize. The standardized products don’t exist yet to address this growing risk; whether it be Liability Insurance, Professional Indemnity, or even Product Liability for the software engineers developing the A.I.

But as more bespoke products are built, and more claims occur, A.I. Insurance will become commonplace fairly quickly.

Free Business Insurance Consultations

If your business is considering using an artificial intelligence tool and would like to understand your exposure, and how to create a bespoke policy for your risk, please Contact Us today.

CCW Global has over a decade of experience in helping businesses in Hong Kong and throughout Asia navigate an ever changing risk landscape and we are happy to offer you and your key stakeholders a comprehensive, no-risk, no-obligation consultation with our expert brokers.