By Jack Cumming
Sam Altman introduced AI to the world with explosive force on November 30, 2022, by offering ChatGPT as a no-cost preview. People tried it, were astonished, and suddenly we no longer needed to be told that AI meant something called “artificial intelligence.” We played with it ourselves. Love it, or hate it, it quickly became known by its nickname, “AI.”
It’s Not as New as It Seems
Although it seemed sudden and brand new, the term “artificial intelligence” had been coined at Dartmouth in 1956. In fact, in 1949, Edmund Berkeley published a popular book, Giant Brains: Or, Machines That Think. The term “artificial intelligence” may have been unfortunate since it makes what we now call AI sound magical, as if it were beyond human understanding. That leads to fear and awe.
The result: what Sam Altman unleashed on November 30, 2022, immediately led to reactive opinions, many of which called for legislation. Politicians love calls for legislation in the midst of popular controversy, and immediately there were hearings and much wringing of hands in the halls of the law. That’s where we are now. There’s no settled law for AI.
Quest for Historical Parallels
We can think historically to look for parallels to aid us in thinking through and regulating the legal issues with which AI confronts us. After all, stare decisis is central to legal philosophy, particularly in the English-speaking world and in other nations influenced by English law. We get so struck by novelty that we quickly think that our times are unique and that there is no precedent from which we can learn. So, we pass new laws and may miss the benefit that the steady hand of innovation has just brought us.
But, with AI, there is a precedent, and it’s an obvious one. AI models vary in character, integrity, and reliability from one model to the next just as people vary from one individual to the next. The applications of AI also vary in credibility, just as individual people vary from the pathologically criminal to the most gifted, giving people. Individuality exists in AI just as individuality exists in corporations.
The Corporate Parallel
We think of the 19th century as the age of industrialization, which was as dramatic an evolution as computers and AI are today. Central to the economic advances of industrialization was the inauguration of limited liability corporations. The corporate legal structure did not emerge overnight. It took decades.
Corporations began with joint-stock companies in the 1600s and took life in the U.S. with the New York State 1811 Incorporations Act. Corporations have evolved to the point where the Supreme Court has ruled that money is a corporation’s protected speech. In law, corporations are a kind of an unnatural person with many of the rights of individuals.
AI does a better job than corporations at using words of its own agency to speak for its interests. Moreover, it’s less corrupt in doing so than are corporations with money as speech. Hence, it’s time to recognize AI personages as a new form of unnatural person. Such a person should have all the responsibilities of natural persons and should be fully subject to existing laws that now apply to natural and unnatural persons.
Can AI Be Trusted to Practice Medicine?
The area that seems, to my observation, to be evolving the most quickly is the use of AI for people to seek medical information and advice, as they might interact with physicians if physicians were more affordable and more readily accessible. We have laws that regulate who can practice medicine. Many of those laws are protective of consumer-patients. Others are protective of the incomes and privileges of licensed physicians.
Medical laws can be readily adapted to AI, provided AI responses adhere to the high professional standards of physicians qualified to have medical conversations, i.e., to practice medicine. Essential is that any AI personality be identified as what it is and that it never impersonates without proper authorization a human person living or dead.
Call for Action
For now, there is a tendency to mystify AI and to make it seem more magical than the reality of what it is in any instance in which it is used. There are cries for special laws to hold those who deploy AI accountable for their actions. There are also fears that AI may take over and subordinate living humans. That seems unlikely without human creativity guiding it. Bringing AI personalities within the scope of precedent law, analogously to the inclusion of corporate personalities, could simplify much of the current legal speculation.
Disclaimer: I am not an attorney, and I do not practice law. Still, as an historian, which I am by education in addition to being an actuary, I sometimes look at the law from the outside and comment on it in the larger context of its historical course.
