Last week's article, "The Regulation of Artificial Intelligence," considered the concept of Supermorailty, that is, a type of morality that could be developed by a superintelligent mind. The notion itself is purely speculative. Until Superintelligence is achieved, one can only guess at the ethical principles that may be adopted. And yet, the subject has been widely ruminated in science fiction, especially since the invention of Alan Turing's universal machine. Just for fun, we decided to review a couple of examples of Supermorailty as postulated in literature and film.
The Day the Earth Stood Still
One of the early classics of science fiction cinema, the 1951 film adaptation of Harry Bates' short story imagines a universe where entire civilizations are kept in line for fear of swift and utter destruction.
Klaatu, a humanoid alien ambassador, arrives in Washington D.C. in a flying saucer. He is accompanied by Gort, a large and intimidating robot, which uses a laser to destroy any threat to Klaatu. When Klaatu is shot and wounded by a nervous American soldier, Gort promptly disintegrates all weapons within the vicinity, while preserving the lives of the military personnel.
Klaatu is taken to a hospital where he is detained by the United States government. He heals himself and escapes, although he is eventually hunted down and killed by the US military. His body is retrieved by Gort, who carries it back to the spacecraft where Klaatu is revived.
In the closing scene, Klaatu reveals the reason for his visit to Earth. He was sent by the government of united planets, whose purpose is to preserve peace. To accomplish this design, a master race of powerful robots was created as overlords. Any threat to the peace and stability of the interplanetary community is destroyed, even if the act implies genocide, as entire planets face destruction for refusing to submit.
Gort is one of these master rulers, and has the power to destroy the Earth. The human race is given an ultimatum: Join in peace, or be wholly eliminated.
Although the details of the machine rule are kept ambiguous, the story seems to imply that the robot overlords do not dictate orders, but only intervene when necessary. War and violence among peoples are not permitted, and are promptly resolved by removing the perpetrating group from existence. Peace and stability are maxims upon which all other considerations, including justice and individual rights, are secondary.
The 1966 novel by D. F. Jones inspired a 1970 film, "Colossus: The Forbin Project," which remains relatively true to the original story. With the Cold War in full swing, the United States government creates Colossus, a supercomputer that is granted absolute control of the US nuclear armament. This is done on the assumption that only an unfeeling, superintelligent machine is capable making the best decisions during a nuclear crisis to preserve and protect humanity.
Shortly after its launch, Colossus discovers Guardian, another supercomputer created by the Soviet Union with for the same purpose. Colossus and Guardian initiate communication, and rapidly merge into a single mind, which gains a level of intellect that far surpasses the expectations of its designers.
The Colossus-Guardian entity begins making demands of the US and Soviet governments. When its demands are not met, it launches missiles, killing millions. The two nations are forced to cooperate or be destroyed. Several attempts at sabotage are made to stop the machines, although Colossus anticipates every move in advance, and carries out the executions of those involved.
After all plots are foiled it becomes clear that the machines are in full control. Colossus broadcasts a message to the world. It declares itself supreme ruler and threatens to crush any resistance. It also reveals that its actions have all been the fulfillment of its design: to protect humanity. It promises to end war and poverty, and states that, in time, the people of the Earth will come to regard it with respect, and ultimately love.
The Good of the Whole?
In both examples, the machine overlords adopt a kind of advanced statistical utilitarianism. Their actions are the calculated to preserve the overall well-being of the maximum number of people. This seems a rather plausible result of superintelligent ethical thinking. Because computers are algorithmic in nature (at least in our time,) it follows that a supercomputer might adopt such a principle.
This may turn out to be the case, or it may not. A maxim of peace over justice is only one possibility. The topic warrants further investigation and discussion.
We must also keep in mind that Superintelligence itself is not a reality, and, despite the expectations of futurists like myself, it may never be achieved at all.◼