Although artificial intelligence (AI) technology has the potential to transform society, the legal issues it raises touch on various areas of law. These areas include data privacy and security, commercial contracts, intellectual property, antitrust, employee benefits and product liability.
AI is generally defined as computer technology capable of simulating human intelligence. Using algorithms, this software can aggregate data, detect patterns, optimize behaviors and make future predictions. Some examples of AI applications include machine learning, natural language processing, artificial neural networks, machine perception, and motion manipulation.
Businesses and organizations often use these technologies to perform functions more efficiently. They can either develop AI capabilities in-house or license AI technology from a third party.
Product Liability Act
AI has become a mainstream feature in many products and services. As a result, the potential for product liability claims has increased. AI’s ability to act autonomously raises new legal questions regarding claims for personal injury and property damage.
Product liability law is largely based on common law and state statutory principles. Claims of negligence, breach of warranty and strict liability constitute the traditional theories of product liability. These traditional theories of accountability also apply in the context of AI.
Negligence claims impose liability on the defendant for failing to meet a standard of reasonable care. This may be because the product was negligently designed or contained inadequate warning labels.
The breach of warranty rights is based on the contractual relationship between the plaintiff and the defendant (the seller of the product). Plaintiff may allege breach of an express warranty or an implied warranty. An implied warranty can take two forms: for the merchantability of the product or for the fitness of the product for a particular purpose.
Finally, strict liability is a standard under which a manufacturer or seller of a product is held liable for bodily injury or property damage, regardless of the level of care exercised.
AI and a bus crash
The application of these product liability principles in the context of AI was highlighted in the recent case Cruz v. Raymond Talmadge d/b/a Calvary Coach. The case featured an AI-based product that has become part of our daily lives: a GPS device. In this case, the plaintiffs were injured when a bus hit an overpass. The plaintiffs’ lawsuit rested on allegations of negligence, breach of warranty and strict liability against the manufacturer of the GPS device. In particular, the plaintiffs pointed to what they characterized as a design flaw in the GPS device. He drove the bus driver under an overpass. The plaintiffs claimed that the GPS device should have been able to discern that the overpass was too low for the bus. The parties eventually reached an agreement.
AI and data privacy
AI technology has created a host of issues related to data privacy and automated decision making. Data protection laws vary around the world. In the United States, laws governing the use of personal data in automated decision-making vary from jurisdiction to jurisdiction. In Europe, the EU General Data Protection Regulationor GDPR, provides a uniform standard.
Despite the differences between the legal regimes, a few underlying principles inform general thinking on these issues. These core principles include the principle of fairness, the principle of purpose specification, and the principle of data minimization.
The principle of fairness requires organizations to treat personal information fairly. This means implementing transparent measures to use personal information within the reasonable expectations of individuals and to mitigate the risk of discriminatory applications.
The principle of purpose specification requires that organizations collect personal information only for specific and defined purposes. This can be difficult to implement in practice because organizations often cannot predict what the algorithms will learn or the correlations the algorithms will make with data sets. This can lead the algorithms to use the data in unintended ways.
The Data Minimization Principle requires organizations to minimize the time they store data and limit the use of data to fulfilling stated processing purposes. This can be difficult to implement in practice, as AI technology tends to perform better with larger datasets.
AI technology and intellectual property issues
AI technology touches on issues of patent law, copyright and trade secrets. The technology is patentable through the designation of class 706 (Data processing: artificial intelligence) in the patent classification system of the US Patent and Trademark Office. Source code and visual elements of AI systems may be protected by copyright law. Finally, trade secret protection can be a useful form of intellectual property protection for AI-related technologies and can apply to algorithms, source code, and AI training datasets.