All eyes on AI coverage risks at PLUS D&O Symposium

Artificial intelligence took the spotlight during the Professional Liability Underwriting Society’s directors and officers liability insurance symposium last week, with experts discussing best practices for underwriting AI risks and potential coverage issues.

D&O coverage concerns regarding AI are currently focused on “AI washing,” or how companies misrepresent their use of the technology and how it impacts their profitability.

Experts who participated in a panel discussion Wednesday said AI can raise privacy concerns if it is created for use with personal data, discrimination and bias risks if used for providing recommendations, and general copyright and IP issues.

“There’s been a tremendous uptick in clients looking for IP-related coverage concerning their AI,” said Nick Reider, deputy D&O product leader (west) for Aon PLC’s financial services group.

Risk profiles for companies either developing or using AI are shaped by what the technology is being used for, with the main risk factors being bias, transparency and accountability, said Claire Davey, London-based head of product innovation and emerging risk at Relm Insurance Ltd.

Ms. Davey said the best way to underwrite risks of emerging technologies like AI is to talk with those developing the technology to understand its capabilities and the governance around it, then get comfortable with the technology and company management.

“Emerging technologies are not uninsurable. It just takes work, effort and research, which takes investment and resources. If we want to remain relevant as an insurance company and as an insurance sector, we have to put in the upfront work with clients, to understand them and provide the solutions,” she said.

When underwriting for a company developing AI, it’s important to look at what the technology does to determine its main exposure and where it can go wrong, Ms. Davey said.

Companies are either proactive or reactive when it comes to AI, and proactive companies will set the standard for what reactive companies do, said Julie Reiser, a Washington-based partner and co-chair of the securities litigation and investor protection practice at Cohen Milstein LLP. She added that recent statistics show that only 14% of corporate boards discuss AI regularly, while 45% have not even discussed it.

“If that’s true in the five-year time frame, that’s a huge problem for those companies,” she said.

Risk exposures must be viewed through the applicable legal and regulatory framework, Mr. Reider said.

Companies focusing specifically on AI must implement self-governance to reduce the risk of lawsuits challenging the directors’ duty of loyalty in overseeing its operations, also known as Caremark cases, he said.

The U.S. has yet to enact any formal regulations for AI.

“Given the regulatory landscape that’s out there right now, additional governance must be in place at these companies. They can’t just be wheeling and dealing willy-nilly with AI,” Mr. Reider said.

The European Union AI Act establishes a framework that categorizes risk levels and prohibits AI systems that threaten people’s safety, livelihood, and rights. The act also expects developers and deployers of AI to exercise care when using the technology and to share feedback with each other.

The panelists said that while AI is already impacting the insurance industry, human involvement is still needed to oversee the technology.

“One of the key governance controls and duties with AI technology is that it does require human oversight. Whilst AI could perform some underwriting stages, you would hope that a human is still reviewing its output,” Ms. Davey said.

AI’s inability to exercise judgment will limit its impact on brokers as they must exercise judgment when working with clients, Mr. Reider said.

The legal industry is benefitting from AI as some creative law firms are using the technology to decrease litigation costs and increase profits.

“Proactive law firms are using AI to reduce the overall cost of the litigation and make more money in the process. That will be an interesting business model trend among law firms over the next two to three years. I expect it to eventually become the new norm, a much more efficient and cost-effective form of litigation,” said Jeffrey Chivers, CEO and co-founder of AI-powered litigation workspace Syllo.