Products incorporating artificial intelligence (AI) can solve many problems faced by old technologies and increase consumer safety. Following rapid technological advances and greater consumer awareness about artificial intelligence (AI), the law governing AI-enabled products is evolving. As companies develop products that implement AI, they should be aware of potential product liability risks and take efforts to mitigate accordingly.
Physical products that incorporate AI such as self-driving cars and medical equipment may expose companies to product liability risks. While calculating every risk a product may pose is difficult enough for the average product, AI-enabled products evolve while in consumer hands. Consequently, those risks and the opportunity for defects continue to grow. Blaming the AI algorithm for new defects is not a defense, even if those defects were not foreseeable. Additionally, the party responsible for alleged defects can be somewhat unclear.
For instance, manufacturers could potentially be held liable for negligence if AI systems are used in foreseeable ways. As an example, take an MRI imaging device that only works well with high quality images. If a consumer is not warned of this requirement and misdiagnoses a patient because it used a low-resolution image, there could be a claim for negligence. Similar negligence claims have already targeted autonomous vehicle manufacturers, and the risk extends to many other industries. What is considered “foreseeable” may evolve as awareness about AI continues to grow.
In strict liability cases, all parties in the product’s chain of distribution (manufacturers, distributors, and retailers) could potentially be held liable for consumer injuries, depending on applicable law. These potential strict liability claims include design defects, manufacturing defects, and failure to warn. For example, if a self-driving car was designed to become better at staying on the road, but this came at the expense of hitting more pedestrians, a design defect claim could arise. Consumers would argue that the algorithm’s designers should have prepared for this issue. In addition, an autonomous car which was accidentally manufactured with outdated software that causes an injury could be a manufacturing defect. Failure to warn is also a significant risk, as potential risks may not be present at the time of the initial sale. Additionally, any assurances about the algorithm’s performance that become incorrect over time could lead to a breach of warranty claim.
A similar claim is “AI washing,” in which companies allegedly exaggerate the amount of AI they incorporate into their products. Misrepresenting the capabilities and functionality of AI in products can lead to claims of misleading investors or trigger SEC investigations. For example, investors recently sued an information technology company, claiming that although the company was advertising its AI-focused operations, it was privately reducing spending on AI.
Mitigating the potential for defects is of vital importance, including assessing potential risks and verifying that the algorithm’s training data is of good quality. Companies can also take other steps to protect themselves from liability.
Transparency with consumers is key. Manufacturers should clearly warn of any potential risks or foreseeable uses of the product in instruction manuals. This includes explaining the device’s accuracy and identifying data with which the product may not perform well. Typical products liability defenses like disregarding manufacturer instructions and misusing the product are available, so greater warnings about how not to use the product can reduce potential liability. Additionally, advertisements should not overstate the capability of these products or a company’s investment in AI.
Currently, there is little regulation specific to AI products. Nevertheless, legislatures and courts will likely begin to shape the law specifically around these products. Therefore, companies should monitor developments in these areas.
Finally, companies should participate in conversations with industry groups and government agencies. Rather than adapting to the law, why not let the law adapt to you? As agencies develop new regulations, they often take industry standards and ethical guidelines into account. Fostering a culture that supports both businesses and consumers can lead to less uncertainty surrounding product liability.
