How will coverages be affected, if at all, by the use of AI?
By Paul Broussard
Artificial intelligence (AI) will without a doubt reshape industries. As AI integration becomes more widely accepted and adopted in various industries, the insurance industry will adapt in response.
Different industries are already using AI tools like ChatGPT and Google’s Gemini to help them become more efficient with manual tasks, but will there be unintended consequences for using such AI for professional purposes? And how soon will the insurance industry react, positively or negatively, to this change?
Implications involving professional liability, media liability/personal injury, and management liability are presumed to be the most affected coverages. Let’s dive in …
- Professional liability. Commonly referred to as errors and omissions (E&O), professional liability intends to cover professionals for their negligence in providing a professional service (usually requiring a license) or for their consultation. Insurance agents, attorneys, and other consultants are starting to utilize AI tools to perform legal research, use written language as a part of their professional work, and to write articles to promote their services.
I have to admit that when deciding to do a topic on this subject, I asked ChatGPT to write this article first to see what it generated; while it was structured well and written eloquently, it lacked a professional standard to pass for a valuable article to my clients and industry colleagues. Having said that, using AI in professional written works could fail to uphold a professional’s standard of care, leading to claims of negligence.
As of now, we haven’t seen any questions on insurance applications asking about a business’s use of AI or AI-related exclusionary language being implemented, but we can only assume it is coming. For example, a lawyer using AI to write a brief that inadvertently cited fake case law is probably something that a professional liability insurance carrier doesn’t wish to cover, because it may not uphold an attorney’s professional standard of care.
On the other hand, technology E&O insurance carriers are
Insureds that are using AI should proactively prepare for the inevitable by establishing written AI practices.
already affirmatively providing coverage for tech companies that are selling AI software as a service. That said, the jury is still out on which way the insurance industry leans regarding AI use in professional industries, but if AI is more widely adopted, insurance companies may be forced to properly address covering these exposures to remain competitive in the market.
- Media liability/personal injury. Media liability intends to provide coverage for claims of defamation/slander/libel, copyright infringement, and plagiarism. Media liability and personal injury coverage parts are typically found in cyber liability and general liability policies.
As I mentioned previously, I had ChatGPT generate its version of this article, and it was an adequate article, but it also contained some plagiarized language found when checked by other AI tools. It contained that language because ChatGPT generates a response based on its learned knowledge of data it was trained on, such as books, articles, and websites.
Generated language could contain use of another’s work or ideas or it may even contain outdated and now potentially false information as facts may change over time. As such, we can expect media liability/personal injury insurance to be the first to implement change regarding AI’s use in the media industry.
It is safe to speculate that plagiarism issues would defy any journalistic integrity that the media industry upholds. So, any exclusionary language that is implemented for the media industry would likely be widely tolerated.
There is already at least one insurer that has a question about AI in their application and they have the ability to use an AI exclusion that applies to the media liability coverage part. For other industries, we do not foresee the personal and advertising injury coverage in a standard general liability policy changing any time soon, which would allow carriers to stay silent on the subject.
While the insurance industry may adapt its coverage for the media industry with AI exclusionary language, change in the insurance industry usually occurs slowly and reactively when the losses start to come in.
- Management liability. This policy can be made up of several coverages, such as directors and officers (D&O) liability, a coverage intended to cover—you guessed it—directors and officers of an organization for their mismanagement that may cause others financial harm. A management liability policy can also contain employment practices liability (EPL) which intends to cover organizations for claims of harassment and discrimination among other allegations.
An organization’s use of AI can cause unintended consequences in both coverage parts.
For instance, there was a recent story that an AI program was appointed as an organization’s CEO by its board of directors. They reasoned that the AI would be able to process data and make strategic decisions better than a human CEO would. However, decisions based on logical algorithms may lack a certain soft skill that doesn’t necessarily work, in theory.
An example would be a large workforce reduction—a culture that was built on the human element could be destroyed overnight if the decision was made purely on numbers. Management decisions carried out without proper discussion surrounding the emotional aspects of a decision could lead to allegations of mismanagement for a board of directors by their investors. Therefore, D&O underwriters may take a harder look at a business relying upon AI to make their management decisions.
Employment practices liability would also be a coverage potentially affected by AI. From using AI to generate images to evaluating investing in a company, individuals and organizations are starting to use AI software exponentially more each day. With that comes greater exposure to discriminatory bias. If an AI system is trained on historical data, it may inadvertently perpetuate biases or unequally represent groups of people.
Using AI to rank or eliminate candidate resumes for job openings, to determine who should get a raise, or to determine who should be promoted will certainly have employment practices liability implications, especially if there is an inherent bias in the results. Likewise, discrimination is not just an employee exposure—allegations could be made from third parties, which would still impact employment practices liability policies with third-party coverage.
Having said all this, we haven’t seen in the wild any AI exclusions implemented on management liability policies, but we could surely see why some may pop up soon.
In conclusion, artificial intelligence will make its way into enough industries where it will force the insurance industry’s hand to adapt to it—whether that is to affirmatively address coverage for it or to exclude it.
Insureds that are using AI should proactively prepare for the inevitable by establishing written AI practices. An AI policy manual would address the intended use of AI in the business, the training on the identified AI programs that are acceptable for use in the business, and the quality assurance/quality control procedures that are in place to provide for necessary oversight.
As AI case law starts to set precedents and shape the legal and insurance landscape, an experienced attorney will likely be a part of your due diligence and risk management. If your insureds are using AI, regardless of any lack of exclusionary language in the insurance industry to date, provide them with consultation to see if there is any current exclusionary language that could reasonably be applied to the use of AI.
The author
Paul Broussard is a risk advisor at Cavignac in San Diego, California. In addition to strategically structuring and placing insurance programs, he is also actively involved in reviewing client contracts to provide a technical insurance perspective so his clients can make informed business decisions. Paul carries the following professional designations: Certified Insurance Counselor, Certified Risk Manager, Management Liability Specialist, Small Business Coverage Specialist, Professional Workers Compensation Advisor, and Cyber Risk Manager.
