The term Artificial Intelligence, or AI, encompasses a wide and seemingly ever evolving range of technology-based products and services. It is employed by everyday products (i.e., Siri and Facebook), and AI cognitive learning platforms, like IBM's Watson, are shaping the future of global industries such as healthcare, automotive and manufacturing. According to research firm IDC, the “widespread adoption of cognitive systems and artificial intelligence across a broad range of industries will drive worldwide revenues from nearly $8 billion in 2016 to more than $47 billion in 2020.”
The rise of AI-enabled devices and robots will create a host of liability issues as they become progressively more autonomous, independent and interconnected. As these robots learn from their actions and take steps farther and farther removed from human programming, there is an increased risk of unpredictability and unforeseen consequences. For companies that employ them, these losses could include, but are not limited to, bodily injury, property damage, and monetary loss. And, autonomous robotic actions could cause a variety of cyber security risks and losses. As a result, companies that employ any type of AI-enabled device or robot should consider the kinds of insurance that could provide coverage for those losses.
Traditionally, robots have been used primarily for assembly-line manufacturing type jobs. But as technology advances they will move into roles that require human interfacing and decision-making abilities. Whether on the manufacturing line or delivering your Amazon purchase, the use of AI and robots will inevitably lead to claims of bodily injury and property damage. If there is a claim against your company by a third-party for bodily injury or property damage a company's first stop should be the standard Commercial General Liability (CGL) policy. Alternatively, if your robot injures your employee, your company should look to its Workers Compensation Insurance for potential coverage.
A CGL policy provides broad defense and indemnity coverage for claims against an insured alleging, among other things, “bodily injury,” and “property damage.” A CGL policy's insuring agreement typically obligates the insurer to “pay those sums that the insured becomes legally obligated to pay as damages because of “bodily injury” or “property damage” (the duty to indemnify). It also typically obligates the insurer to provide and/or pay the litigation costs related to that claim, including the attorneys' fees incurred in defending a lawsuit (the duty to defend). The “bodily injury” or “property dam-age” must be caused by an “occurrence,” which is usually defined as, “an accident, including continuous or repeated exposure to substantially the same general harmful conditions.”
Based on the bare language of the insuring agreement, a CGL policy should provide insurance coverage for “bodily injury” or “property damage” caused by an AI-enabled device or robot to a third-party. If an insurer wants to limit this coverage, the burden is on it to prove that an exclusion in the policy applies. While an insurer could argue that one of the standard GCL exclusions applies, currently there are no AI related exclusions for “bodily injury” or property damage.”
AI-enabled devices and robots are also permeating the professional world and performing roles once reserved for humans. As these entities increasingly provide services traditionally performed by humans, including medical and legal services, companies could face professional negligence claims that arise out of their reliance on AI technology. As a result, companies should consider their Errors and Omissions (E&O) insurance. E&O insurance, often referred to as malpractice insurance, is intended to insure against liability arising out of an act, error, or omission of the insured in rendering or failing to render services. While there is no “standard” wording for E&O coverage, most E&O policies provide insurance for judgments, settlements and defense costs arising from professional liability claims of negligence, misrepresentation, violation of good faith and fair dealing, and inaccurate advice. 7A Couch on Insurance (Third) §103.3 (2013).
Specialized E&O policies are also available for technology-related professional liability claims. Tech E&O polices, while varying greatly in form, generally provide coverage to tech providers for losses resulting from technology services, technology products, and network security breaches. As AI agents and robots perform progressively more autonomous professional services, an increase in specialized E&O policies and insurance products is likely. For example, Lexington Insurance, a subsidiary of American International Group, recently introduced ”Robotics Shield,” a policy suite that combines “general liability and product liability insurance, robotics errors and omissions insurance, and specialized risk management, all specifically tailored to the robotics industry.” Based on the proliferation of specialized cyber policies in the last decade, Robotics Shield is probably the first of many policy forms that will address AI and robotics.
AI-enabled devices and robots could also perform autonomous cyber actions against others, including an insured's own company, a client, or an anonymous third-party. Companies face considerable hurdles to obtain coverage for third-party cyber claims under the CGL policy. Since 2014 standard CGL policy forms exclude coverage for “bodily injury” arising out of a data breach. In addition, insurers have taken pains to draft language that data breaches and loss do not constitute property damage. For example, in 2001 the Insurance Services Office (ISO) amended its CGL policy form to clarify that “electronic data is not tangible property.”
The CGL policy also provides defense and indemnity coverage for claims against an insured alleging “personal and advertising injury.” There has been limited and divergent case law on whether losses from cyber security events are covered under the “personal and advertising injury” clause of the CGL policy. See, e.g., Zurich Am. Ins. v. Sony Corp. of Am., No. 651982/2011 (N.Y. Sup. Ct. Feb. 21, 2014) (while a hackers' theft of confidential data on tens of millions of Sony PlayStation users constituted a publication of private information, the claim was not covered because the cyber information was published by the hacker, not the insured); Travelers Indem. Co. of Am. v. Portal Health Solutions, No. 14-1944 (4th Cir. April 11, 2016) (a computer server failure that resulted in the disclosure of medical records was a “publication” and the insurer had a duty to defend under the personal and advertising injury coverage provision). In part because of these limitations in the CGL policy, there are now a wide range of specialized cyber policies that purport to cover data breaches and cyber events. Therefore, if a company has purchased a cyber policy it should review for potential coverage if an AI-related cyber loss occurs.
To maximize the available insurance coverage for AI related losses, a company should review and understand how and for what purposes it uses AI technology. In combination with this review, insureds should review what risks and claims they currently do or could suffer. Insureds should also re-view their CGL, E&O, and any specialized tech E&O and cyber policies. And, perhaps most importantly with any emerging technology, companies should not make any assumptions about what is covered under their policies, let alone assume that an insurer is right in disputing coverage.
Originally published on newyorklawjournal.com. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.