As humans we have the unique qualities of consciousness and emotional intelligence; the capacity for empathy and compassion; and the feelings of love, disgust, queasiness, and hate. We have evolved these emotions along with our sense of self-awareness, and they won’t likely be replicated anytime soon by artificial intelligence technology.
Some experiments have shown that our capacity for empathy and compassion is a product of mirror neurons that enable us to read and translate the feelings of others. Those without this emotional feature are often diagnosed with autism. But the more extreme examples are those recognized as sociopaths and psychopaths, who are unable to experience the distress or pain they inflict on others. Thus the moral guideline of the golden rule, reciprocity, is only an abstraction to them.
In her book, Girl Decoded, Rana el Kaliouby explores the scientist’s quest to reclaim our humanity by bringing emotional intelligence to technology. Her company uses photographs of human nonverbal behavior (i.e., facial expressions, body language) to correlate those behaviors with self-reported emotions in order to identify patterns of reliable, visible emotional expressions. Thus, for example, the feeling of disgust may be universally recognized by particular facial expressions regardless of one’s culture, race, or family experience. This expression may be observed soon after birth as a genetic predisposition. Kaliouby’s company has developed technology that translates these nonverbal cues into language for autistic children. These children can wear devices that inform them of other people’s emotions so that they can interact with other children more effectively.
Most scientists that work in the field of artificial intelligence are doubtful whether machines will ever have the capacity to model human consciousness or self-awareness. However, computers like the ones that Kaliouby creates may someday be manufactured as robots with artificial faces that can model human emotions that replicate sadness, happiness, love, hate, and queasiness. They may be able to speak words that contain emotional meanings such as “Oh, what a beautiful morning” or “I feel happy today.”
The Turing test for artificial intelligence, originally Alan Turing’s “imitation game” from 1950, starts with a black box. A person on the outside of the box can converse with the box but isn’t able to tell whether a person or an AI machine is on the inside of the box by asking it questions. Thus the Turing test for artificial intelligence may successfully be met by future generations of emotional intelligence machines that can be asked, “How do you feel today?” The answer the machine gives may be relevant to the question without the AI itself experiencing feelings. Thus, fooling the questioner into believing that a person is in the box.
The purpose of this article is to inform the reader that we, as humans, have a distinct quality of emotional intelligence that uniquely enables us to assess what we collectively regard as “fairness.” Fairness is an emotional response to a complex set of facts. It’s doubtful that artificial intelligence as we know it will be capable of judging fairness. Fairness, like morality, is a result of an evolutionary process that is a product of interaction in a social network. Our judgment of what is fair can be altered by the insights of others, causing us to reevaluate our perspective. Courtroom judges historically apply the common law as a tool for resolving moral conflicts with the purpose of reaching a fair resolution among competing demands.
On the other hand, AI machines, like umpires, can judge observations about whether to call a ball or strike based on the strike zone of the batter. Emotions are irrelevant to the judgment of whether a ball throw is a strike or not. Also, AI can determine more accurately than a human whether a spot on an x-ray is a sign of cancer.
Judges play a different role in our society. Their decisions are intended to convince observers that their call was fair. Judges are responsible in their decisions to win the public’s confidence that everyone will receive a fair hearing. Otherwise, people with conflicts will be incentivized to take the law into their own hands if they don’t believe that their grievance will be adjudicated fairly. This means that the judges must demonstrate in their decision that they took seriously the arguments presented by the attorneys who may have lost their case. The judges must show losing lawyers that their cases weren’t supported by the facts or the precedents that have been applied uniformly over time. Thus judges’ decisions need to be seen as objective rather than reflecting personal prejudice. There are usually many precedents in judicial decisions that are relevant to the outcome. Judges demonstrate in well-written decisions that the most compelling precedents favored the outcome adjudicated. People who read a judge’s decision will be influenced to engage in conduct consistent with fairness under similar circumstances. Thus a judge’s decision provides a model for moral behavior.
Judges, however, are human beings who are emotionally affected by narratives that the adversaries present to the court. It’s unnatural to expect judges not to be so affected by their sense of fairness. The nature of democracy is based on a belief that adversaries can argue over their conflicts and reconcile their differences through discourse by compromise or adjudication according to fair rules. In litigation these fair rules of procedure are referred to as due process. In informal discourse these fair rules of discourse are referred to as “reason’s fulcrum.”
Reason’s fulcrum means that when parties have a dispute with each other, there is an implied understanding that each party will concede to the wisdom of the other if she or he has a better or fairer solution. If either party is unwilling to have a change of mind when presented with evidence to the contrary, then discourse is futile. The benefit of discourse is that each party has a different life experience or genetic makeup, or facts or ideas unknown to the other party. Therefore, if one of the parties is unwilling to consider the other’s argument because she or he is speaking for God, or referencing books written by God or people who claim to be God’s agents, such as popes or kings, then dialogue is useless. An argument based on God’s authority is a conversation stopper. It is another way of saying “My way or the highway.” It is undemocratic because it violates the terms of “We the People” in the preamble to our constitution that presumes that the collective conscience of people can create a better society in service of the general welfare. Our preamble presupposes that people have the agency to make ethical and policy judgements based on the exchange of life experience.
AI machines, on the other hand, at least as we presently know them, will likely never have the capacity for social interaction to be influenced by changing social contexts and evolving emotional responses to moral arguments.
The Relationship of Religion to Morality and Public Policy
A religious worldview, like artificial intelligence, is similarly not equipped to address the democratic, secular moral, or legal enterprise, but for different reasons.
The religious worldview of the Abrahamic religions (Judaism, Christianity, and Islam) supposes that a belief fixed in ancient times is the foundation for right action. In a theocracy it concludes that religion is the necessary foundation of good government. It isn’t dynamic and doesn’t entail a process of learning from life experience, constructive debate, and the discoveries of the social sciences about the nature of humans.
Radical philosophers, on the other hand, say that contextual, contemporary understanding, not belief, is the source of morality and that a reasonable government embodies this understanding in its laws. In other words, religion is deductive whereas democracy and science are inductive. Religion is hierarchical with its source of morality, God, as irrefutable. Democracy, morality, and law are inductive. They evolve as human minds, language, and science evolves. Inductive thinking begins with the recognition of patterns in human observations that results in testable moral assertions. The probity of conflicting moral claims is based on the accuracy of relevant facts and the legitimacy of precedents that have applied in the past to similar conflicts. There are better and worse solutions to moral conflicts, just as there are more and less fair judicial decisions depending on the resonance with their audience. Thus, decisions perceived as partisan undermine the public confidence in our judiciary.
Human rights laws are self-evident based on the emotional appeal that strikes a chord within us. It’s convincing when we feel aggrieved by their violation. The radical philosophy of our founders says that observations from experience, not faith, are the foundation of morality, and the public good is the private good perspicuously understood. Belief in God, the soul, or other invisible authorities is irrelevant to discourse regarding what constitutes good government. Similarly, claims of God’s dictates aren’t persuasive in our courtrooms or legislative debates. Legal disputes about personal liability, criminality, or contracts are resolved without reference to religious authority. Thus, invisible authorities of religion are incapable of democratic dialogue because there is no room for reason’s fulcrum to persuade based on evolving attitudes of fairness.
Our Declaration of Independence was a statement of grievances against Great Britain that were so egregious that they violated our dignity of what it meant to be human. The statement of those grievances was sufficient to resonate with the colonists’ emotions sufficiently for them to sacrifice their lives at war to dissolve their political bonds with Great Britain. These grievances were statements of natural human rights, not divine rights of God. They were products of collective human emotional intelligence.
The introductory phrase of the Declaration of Independence—“when in the course of human events”—alerts us that the event to be announced doesn’t arise from any divine dictate. The colonists recognized their natural powers to repudiate a colonial government because of reasons disparaging of their human dignity. It makes clear that the reasons for revolt are circumscribed by experience of this world, not sins dictated by a disembodied spirit.
Religion is hierarchical with its source of morality, God, as irrefutable. Democracy, morality, and law are inductive. They evolve as human minds, language, and science evolves.
In Thomas Jefferson’s original draft of the Declaration of Independence, the truths stated there were described as “sacred and undeniable” before they were changed to “self-evident.” This change served to alert us that there will be no appeal to truths handed down from scriptures or announced by prophets. There will be reasons for beliefs, not beliefs without reasons. We “hold” these truths to be self-evident because our free understanding is an active power, not a passive one. It doesn’t consist in the capricious affirmation or negation of random propositions. Its job is to realize our freedom by rendering the unintelligible intelligible, and succeeds when it holds things up in such a way as to resonate with our understanding of moral truth. The opening of the Declaration doesn’t guarantee that the claims that follow are true, but that they can be demonstrated to be true by moral persuasion. We are equal and have unalienable rights not because we take on faith that we represent an exception to the laws of nature but because we understand that we are a part of nature and subject to the same universal laws that make all things explicable.
The Bill of Rights is a statement of natural rights of individuals that were believed to justifiably be insulated from government interference and protected by the constitution. They were not divine rights—nor were they based on religious scripture or sourced in God. The anti-federalists were so committed to the Bill of Rights that they wouldn’t agree to ratify the US Constitution without their inclusion.
There is nothing in religious scripture or claims by people who regard themselves as God’s agents (i.e., popes or kings) that dictate the natural right of humans to “life, liberty, and the pursuit of happiness.” In fact, Judaism, Christianity, and Islam all revere the biblical lesson of original sin that Adam’s disobedience by seeking the metaphorical pleasure of the search for knowledge (symbolized by the appeal of the apple) is the burden all humans must bear. Also, redemptive suffering is honored in Christianity for imitating the martyrdom of Christ. Finally, Christianity disparages happiness in this life, felicitudo, in favor of beatitudo, happiness in heaven. Jesus did not repudiate either slavery or torture, the most horrific human concepts in human history, although Jesus was a victim of torture.
On the other hand, American values are distinct from religious values for repudiating the divine right of kings or of any divine source for our governance. The right to human fulfillment and to own property are not Christian values. Our founders were influenced by Enlightenment values, which provided human criteria for judging public fairness—the collective conscience of We the People.
In conclusion, artificial intelligence as we know it, based as it is on machine learning, is unlikely to have the capacity to replicate the collective conscience of human judgment in resolving moral conflict. And the Abrahamic religions function by appealing to faith, which exempts their moral or truth claims from any criteria of falsification. This violates reason’s fulcrum and democracy’s foundation, a morality built upon social dialogue.
A secular perspective on morality is equivalent to the collective conscience through social discourse, in pursuit of a better society in service of the general welfare. Like the preamble to our constitution, it’s a moral process that is a social enterprise, not an individual pursuit. It’s a product of the Enlightenment.
Our common law, the law that applies to civil disputes, is an excellent model for secular morality. It doesn’t apply supernatural ideas, such as the authority of God, in search of a fair resolution of moral disputes. It involves rules for fair engagement, such as due process, rules of evidence, and precedents that have met the criteria of fairness over time. If lawyers can convince a judge that there are better precedents applicable to a case, then the judge can adopt new precedents that reflect a fairer outcome. This is how the common law evolved.
There are many moral questions that aren’t justiciable, but the model that the common law offers is our best guide for moral decision making. This is the basis for secular morality and the moral foundation of humanism.