Ever wonder why it’s called Artificial Intelligence?
Every day it seems there’s another article on artificial intelligence technologies. I’ve been reading about ChatGPT. It uses machine learning to find patterns of patterns in training data, mostly written by humans, to produce human-sounding prose in response to prompts. I recently experienced this secondhand from our son Stephen.
Stephen turns 36 today. He’s a public high school teacher—loves his kids, they love him. A few weeks ago he emailed ChatGPT asking it to recommend charitable projects that his National Honor Society students could organize in school. Stephen told me he received a reply in 15 seconds. It was a neatly laid out memo featuring eight ideas, all sounding like they were written by an intelligent human being. What’s so bad about that?
Nothing—unless you’re a dinosaur. C. S. Lewis introduced himself as one in his inaugural lecture at Cambridge in 1954. He meant he was a native of the medieval world, an age that he felt was becoming extinct, as dinosaurs had long ago. Because of this, Lewis recognized the Middle Ages defined “intelligence” and “artificial” differently than most do today.
The term intelligence comes from the Latin intellectus, meaning to comprehend or perceive. In the Middle Ages, it was “understanding in which truth offers itself like a landscape.”[1] Genuine truth unfolds naturally, gradually over days, weeks, months, even years. The Truth must dazzle gradually / Or every man be blind. Blind to what? To God as the source of wisdom and understanding in which truth offers itself gradually through contemplation.
A contemplative life is necessary for we cannot know God the way we know anything else. He is infinite, everything else is finite. The degree of difference between the two is, well, infinite. Knowing God’s infinite truth requires contemplation which is patient and passive. It doesn’t come by typing a request that algorithms immediately spit out in human-sounding prose. That’s called artificial intelligence, or, in this case, the artifice of intelligence.
The term “artificial” come from the Latin artificium, meaning “handicraft,” something made by human beings rather than occurring naturally. That can be good or bad. Robotic surgery is a good A.I. technology. Artificial Intelligence technologies are bad when they become an artifice, which means contrived or false. The artifice of intelligence makes people “see only what new technologies can do and are incapable of imagining what they will undo.”[2] My sense is wisdom is being undone, reminiscent of T. S. Eliot’s fear: “Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?”
This is what Dinosaurs fear. They feel ChatGPT raises the odds of students cheating on their homework. They feel DALL-E 2, an A.I. technology that lets you create digital images simply by describing what you want to see can become generators of false images. They appreciate how GPT-3, an A.I. language system, can write, argue and code. But dinosaurs fear GPT-3’s implications for the future could be profound.
I don’t know how many of Stephen’s students will sense this. The latest Pew study shows the decline of Christianity and rise of religious “nones” continues, especially among young people where it appears to be accelerating. The damage is that “nones” will find it increasingly difficult to imagine natural intelligence is wisdom coming from a Real Person, God. You kidding me? Dinosaurs like Lewis and Eliot were not kidding.
Nor is Jonathan Haidt. He sees the damage being done to today’s young people. It too is due to the Artifice of Intelligence. We’ll consider Haidt’s research next week.
[1] Josef Pieper, Leisure: The Basis of Culture (Ignatius, 1952), 28.
[2] Neil Postman, Technopoly: The Surrender of Culture to Technology (Random House, 1993), 5.
You might find Non-Computable You by Robert Marks to be an outsiders take on AI
Thanks for the post, Mike. Your point and use of the label “Dinosaur” is great. It makes me think of the Jurassic Park syndrome which describes scientists being so bent on an exploration to see if they *can* do something without actually asking if they *should* do it. There are a lot of ethical questions regarding large neural network models, including ChatGPT, and the types of things they could – or, more restrictively – should, be used for. Being a Dinosaur seems on point to me.
“NON-COMPUTABLE YOU” by Professor Robert J. Marks, as recommended above. Prof. Worth a read. Marks is an AI pioneer, having founded the first, or one of the very first, peer-reviewed AI Journals. Yet he is a dissident from the “brave new world” continuously paraded before our eyes in the form of recycled sales pitches and sales hype (Professor Marks has a test for detecting sales hype in his book). One of his key convictions would appear that we bear the image of God, and that machines are unable to replicate that – for example, they do not do well with what is non-algorithmic. In my own case, I work as a conference interpreter (a translator of meetings). The vastly greater part of my work is non-algorithmic and the machine is not able to compete head-to-head with what I do. Another example, recently I spoke with a man who spoke to an ophthalmologists’ conference recently about exactly why AI technologies cannot replace those who care for our eyes, as the eye is specifically not a camera (among other reasons derived from the relentless onslaught of AI hype). Yet, developers of A.I. (artificial intelligence produces artificial translations) are continually dependent on recycling the “amazing progress” they are continually making. At its best, AI may work in many field as an “exoskeleton” but nothing more except in many or most cases. Unless we fail to think. As Mike suggests, what will we lose until such time as people figure that out? Perhaps it is the case that Christians are again called to live counter-culturally?