Legal issues accompany the benefits of using AI

In the 1968 film, "2001: A Space Odyssey," HAL 9000 was a computer that controlled the systems of the spacecraft. HAL terrorized the spaceship's crew with a soft, calm voice. HAL was created by humans and then worked to destroy his creators. The astronauts attempted to conceal what they were saying but were unaware that HAL could read their lips. HAL decided to kill the astronauts in order to protect himself and continue with his evil program.

HAL 9000 is one of the many images come to mind when the term "artificial intelligence" is mentioned. The term can mean different things to different people, including images of self-driving cars, human-like robots and drones. In fact, AI is a very broad term that refers to a multitude of different concepts. In many cases, AI performs frequent, high-volume, computerized tasks performed reliably and without stress. AI works by combining large amounts of data with extremely quick processing that allows the software to discern patterns and to organize such patters to predict future actions. For example, AI can be used to process millions of photographs so that a computer can estimate the location where a certain picture was taken. AI has also been used to analyze all of Rembrandt's known paintings and to create its own Rembrandt facsimile painting, even including such things as the typical paint thickness used by the artist. Conversely, AI can also refer to technology as simple as a card key-coded entry door or a vending machine that dispenses food. In a sense, AI is simply the use of technology to easily perform tasks that could otherwise be performed by humans.

There are some significant legal policy issues raised by the emergence of AI technology as it permeates our daily lives. Some of the issues that are generating much discussion involve intellectual property ownership, as well as privacy and liability considerations.

IP Ownership

AI can be used to produce novels and paintings, and to develop cutting-edge inventions all on its own. Who will own these inventions that would otherwise be protectable under our patent, copyright and trademark laws? Some have suggested that a separate type of IP registration should be instituted for inventions developed through AI. Many consequential issues likely will arise, such as the possible rights of software engineers who create the AI to own part of the works created by their software.

U.S. copyright laws protect works that are developed through an author's skills, work and exercise of decision-making. Would works created through AI be entitled to protection under this standard? Case law suggests that an author must be a human. In the recent Naruto v. Slater case (888 F. 3d 418 (9th Cir. 2018), the 9th U.S. Circuit Court of Appeals found that a monkey lacked standing to sue for infringement of his selfie. Registration of a patent requires that the invention is new, useful and not obvious. Would this apply to inventions created through AI when the process of development by AI naturally involves the analysis of millions of items of prior art?

It is clear that existing law lacks mechanisms for identifying AI as an inventor. IP laws generally require a rights holder to be either an individual or a recognized business entity. The rapid development of technology means that AI will soon lead to new discoveries and inventions that should be protected. Some legal framework to address this ownership problem must be available soon, and it is the responsibility of humans to develop these laws. This is one matter that cannot be left to the machines.

Use of Data

Serious privacy and equal treatment issues arise out of the methods that businesses use to analyze and capitalize on the huge amounts of data generated by smart devices and other processes that have become increasingly computerized and electronic.

For example, the use of data analytics in the insurance industry creates a high risk that customers' private information will be used to evaluate insurability and rates. Lending institutions, or almost any company that deals with a high volume of business, often encounter the same problems. Customers want to be assured that personal data is not monetized as a commodity and that they can effectively control how their data is used.

AI also has the potential to use data generated by self-driving automobiles to determine where one has been, predict where one is going and whether any traffic laws have been broken. Such data could also be of potential value to insurance companies wishing to make a low risk assessment of potential insureds. More serious consequences would follow from insurance carriers' possible similar uses of genetic data, which might predict certain diseases and therefore impact the cost of health and life insurance. Already, new companies are working on technology that uses digital images to examine an individual's physical features to determine longevity, health status and disease susceptibility.

The European Union's General Data Protection Regulation (GDPR) places restrictions on the use of AI to evaluate private data to make decisions. Likewise, the California Consumer Privacy Act of 2018 (Senate Bill 1121) follows in the footsteps of the GDPR and regulates the use of personal information by many businesses. The act will become effective on Jan. 1, 2020. These are just the first of a multitude of regulations that will likely arise out of the proliferation of AI in business decision-making.

Liability Issues

Another significant issue that will arise out of the proliferation of AI involves liability. As an example, current laws provide that the driver of an automobile is liable for traffic infractions and criminal violations, while the manufacturer is potentially liable for vehicle defects. The use of AI in developing autonomous vehicles has the potential to mix these two concepts. Thus, if an autonomous car is in an accident and the "driver" has not exerted any control over the vehicle, it may be said that the accident was caused by a vehicle/software defect that is the responsibility of the manufacturer. Presumably, the software should be able to instruct the vehicle to obey all speed limits, stop signs, etc. Machines can make mistakes. Thus, any crash would be the computer's fault.

However, the law would also need to address the scenario where the "driver" is able to disengage the autonomous functions and take over control of the vehicle. In this case, would the "driver" be responsible for a crash if he or she unsuccessfully attempted to intervene?

As AI is being developed for use in almost every facet of American business, it has the potential to impact everyone that a person hires, including professionals such as doctors, lawyers, accountants and so on. Already, AI is being developed for use in each of these professions. Regulatory agencies will have to develop new regulations to address liabilities that may arise from the professional's possible abdication of responsibility to a machine.

Conclusion

AI offers to do a lot of our work for us. However, there are many legal issues that accompany this benefit. As HAL 9000 said: "I am putting myself to the fullest possible use, which is all I think that any conscious entity can ever hope to do." Humans must do the rest. 