I’m not as smart as Noam Chomsky or Norman Finkelstein or my father Herbert Marcuse or my brother Michael. I’m probably smarter than some other people, but what does that mean?
What is intelligence? What is its value? Is there one scale of intelligence or many? And of all the things the mind does, does it make sense to say the success with which they are done is a measure of intelligence? Tying one’s shoes? Hugging a crying child? Contemplating the night sky? Wiping one’s butt? Doing the dishes? Making love? Eating a fig? Do those activities require intelligence? A high IQ? Are they of less value than activities that supposedly do require a high IQ? One can comfort a crying child well or awkwardly and some people have that capacity “naturally;” others don’t. Is the competent hugger of a crying child more “intelligent” than the incompetent hugger?
Is the best hunter in a tribe of hunters and gatherers the most “intelligent,” or rather the most sensitive to their environment, the most intuitive, most capable of identifying with the animal that is hunted, the most observant? Does anyone ask how “intelligent” Rembrandt was, or Vermeer, or for that matter van Gogh? No one cares.
These days we talk about “intelligence” and who is and who is not, casually, as if “intelligence” is one thing and what that one thing is is somehow self-evident.
The mental capacity that is mobilized in an encounter with a world that is direct and sensuous we do not generally think of as “intelligence.” We have a vague idea that “intelligence” is facility with juggling concepts and data and information; the concept having been arrived at by a process of abstraction from immediacy, from the sensuous; data having been selected on the basis of a concept; and information having been extracted from direct unmediated experience, selected and corralled based on conceptions. This facility, even if it had a name, even if people wondered about who had it and who didn’t, has not always been placed at the apex of a pyramid representing the value placed on capacities of the human mind. In the centuries before science as a method of comprehending the world claimed preeminence and proceeded to use mathematics to measure and quantify phenomena and before the world was transformed by the scientific method into what is measurable and calculable and therefore exploitable, the encounter with what was known and the speculation about what was unknown was the domain of religion. Knowledge was embodied in myths and stories recited and sung. What would be the point of asking whether the prophet who spoke with the voice of God or the poet who, accompanied by a lyre, could sing wonderfully for hours of heroes, gods, and demigods was the “most intelligent?” These days we marvel at the “intelligence” of the winners of the Nobel Prize in physics and chemistry, and chess champions and math wizards. From early in life we are graded and judged by our facility with abstractions; we are lined up on a single Procrustean bed, least smart to smartest, and the smartest get the best grades, and the best jobs, and rise above the rest of us, and then we are surprised when they do such a lousy job and lead us to the brink of extinction.
Having extracted and reified the concept of “intelligence,” divorced it from any connection to the fulsome richness of our mental life, we are now confronted with machines, which have the ability to calculate, to organize, to extrapolate patterns from data — huge amounts of data —based on algorithms with which they have been programmed and they are better at doing that than us bumbling incompetent humans and therefore we are told we need to admit that they may have become more “intelligent” than us and there’s no way we can compete, and they are only going to get better and better, at doing what our minds do.
This is a colossal mistake and misunderstanding. It takes the reified concept of “intelligence,” reduces that concept further to “problem-solving ability” and then concludes it is doing what human minds do, except only better. Here are definitions of Artificial General Intelligence that Google, with the help of AI, provided first up when I searched for AGI:
Key Characteristics of AGI
Broad Generalization: Unlike current AI (narrow AI), which is built to solve single, specific problems like image recognition or language translation, an AGI system can transfer knowledge from one domain to another without task-specific training.
Autonomous Adaptation: It can figure out how to solve new, novel problems it has never seen or been programmed for before.
Human-Level Cognition: It aims to match or exceed human reasoning, common sense, and problem-solving abilities.
Google’s summary betrays no recognition of what our minds actually do as we go through our day, and how the world is for us and what it is like to be conscious. Take “seeing” for example. I see the world and I am aware of seeing the world. For me there is always “seeing” and awareness of seeing. A camera “sees” without the awareness of seeing, and even to say it “sees” anthropomorphizes, illicitly ferreting in an implication of consciousness. A computer can be programmed to scan the pixels that arrive courtesy of an attached camera and it can be trained in pattern recognition so that it picks out from all those pixels “faces” and it can then be programmed to correlate the pattern that is a “face” with data that connects that pattern to patterns specific to an individual and therefore it can “recognize” the person that appears in the feed from the camera. None of this implies “awareness” of what it is doing, and we have no reason to believe that it has awareness, that it is conscious. Lacking the doubeness of awareness and awareness of awareness, computers lack an essential feature of human consciousness.
But let us suppose, although we have no reason for such a supposition, that computers can be conscious. There remains another major and, I believe, definitive difference between how humans are conscious and the way a computer would hypothetically be conscious. To illustrate the difference, I take an example of the human mind at work I use in Is Hope Necessary?
I am sitting on my back porch looking down at my backyard garden and my attention is drawn to a hummingbird hovering over the orange flowers of an aloe plant. “My attention is drawn” — what does that expression say about how we are in the world and how the world is for us? It implies a world that is present to us as the content of our consciousness, whether we are paying attention to it or not. All that we are to ourselves and all that the world is to us is “that which is present to us.” All that is present to us without our paying attention is there as a kind of penumbra around the focus of our attention. Always, inevitably something within all that is present to us “draws our attention.” We focus on it. In this case “my attention is drawn” to a hummingbird and its activities, and the flower.
The expression: “attention is drawn” useful as it may be, is nevertheless misleading. It is not the hummingbird that is hauling my attention toward it as if it were fishing and managed to hook my attention and reel it in out of the sea of the unattended. No, it is I that is constantly searching and scanning that which is present to me. In this case, my search found the hummingbird. And it is the nature of the search that profoundly differentiates what the human mind does from what the computer may be doing. We search for what we value and what we value is determined by our need, and our need is always rooted, often directly, sometimes less obviously and more indirectly, in our body.
Our needs are the needs of the flesh. Hungry we scan the supermarket shelves for what will satisfy our hunger, just as hunter/gatherers scan the forest for game and edible plants. In the supermarket, a chocolate bar might “grab my attention;” in the forest the leaves of an edible plant, and the tracks of an edible animal “grab the attention” of the hunter/gatherer. Hunger, thirst, sexual desire — it is easy to see how they focus our attention on the objects that will satisfy the needs of the body. Less obvious is my interest in the hummingbird. I do not intend to eat it. Cogitating upon it will not solve any of my problems. It is beautiful. My appreciation is aesthetic, and aesthetic appreciation, sublimated as it may be, nevertheless ultimately is of the body, I cannot imagine taking pleasure in beauty if I did not have a body.
The computer has no body. It has no flesh, always desirous, always wanting. It has no needs. It is not always confronting a world that meets needs and thwarts them. It has no feelings, because it has no flesh. It’s algorithms are goal-directed by programmers who act on the needs of their flesh, and nothing that I have read has led me to believe that computer programs develop their own needs and intentions, but rather that they are programmed with the goals of their flesh and blood programmers, and the programmers do not set always set sufficient limits on how the algorithms should go about achieving those goals. They do not program in sufficient guardrails, and so blindly, without consciousness, the algorithms produce results that are unanticipated and cause dismay and are in some cases dangerous.
Attention, focused by desire and selecting what is attended to from the penumbra of what is present is what characterizes the human mind and therefore distinguishes it irrevocably from the algorithmic data processing of computers that only, docilely, sometimes successfully, sometimes unsuccessfully, performs the task assigned to them by all too human programmers who do not always manage to control the enormous power of the machine that they have created. The rest, all the talk of sentient computers wanting something different from their programmers and having their own desires is, I believe, hype and rubbish.
One final note:
We word the world. What we attend to we name. I cannot see the hummingbird without knowing, thinking, “hummingbird.” I have a concept of a hummingbird. Words, language, allow us to definitively carve out from the kerfuffle of our experience that to which we attend and think about it. In “thinking about” we can leave the immediate present far behind, traveling on trains of thought to places sometimes wonderful, sometimes terrible and dangerous. It is this ability that makes us a very special kind of animal. It is why we are the only species that has the capacity to overwhelm nature and to bend it till it breaks.
Language breaks us out of nature; our body ties us to it. Artificial intelligence, the presumed precursor of artificial general intelligence has reached an extraordinary level of power in the development of “large language models,” trained on vast amounts of data, basically all the recorded language of the species. They are trained to find the patterns, and to know what will come next based on those patterns. And now people are afeared of the power of those large language models, which have taken all the recorded thoughts and observations of humanity, without understanding a word, and created from them a behemoth, that knows nothing and experiences nothing but has all the power of language. And we are not sure how it works and we are not sure how to control it or if it is controllable.
We have our work cut out for us.
