Artificial Intelligence: Garbage In, Garbage Out? New Oxford Review.
Christopher M. Reilly, Fredericksburg, Virginia,
New Oxford Review, Letters. November 2024.
I greatly appreciated Bob Weil’s article “Wrestling for Truth with ChatGPT” (Sept.). He describes the inherent flaws of generative artificial intelligence (AI) with a deft hand. He also suggests a crucial element of responsibly managing our relationship with AI: We must approach it with “critical, well-formed questions and some prior knowledge of the subject.” I agree that we need to replace our existential anxiety about some kind of AI apocalypse with a wiser trust in God’s providence.
It is important to recognize that Mr. Weil mostly refers to a specific kind of technology called generative AI, which currently takes the form of “large language models” like ChatGPT. This is a brand-new technology that will be refined and improved substantially in just a few years, and yet it is also a fundamentally flawed approach to a broader machine intelligence — at least without the integration of other technologies. Criticizing generative AI in its current form is a bit like chastising Thomas Edison for a flickering lightbulb 30 minutes after its invention. But Edison never made wild claims — comparable to the hype around AI — that his lightbulbs would replace the sun!
The core problem with AI is its structural tendency to motivate sin. Whether it is massive theft of data, development of a corporate surveillance nightmare, facilitation of child pornography, or crass imitation of the unique emotional and rational dignity of human beings, AI strongly encourages a utilitarian and power-based disposition toward the world and other persons. It has the overwhelming potential to saturate our culture and suppress or distort human relationships. We have the right and the duty to say “no.”
______
Karl Stephan of San Marcos, Texas begins,
“Bob Weil’s experiment, in which he asked ChatGPT controversial questions and got back the standard liberal answers, is not entirely surprising. But it is troubling, because it is indeed uncertain at this time “whether we are training AI or it is training us,” as Weil writes…”
Catholic Critique 