On September 7, 2025, The Wall Street Journal ran a column by Greg Ip. Here's a link to the column, which was headlined, "AI Risks Choking Off New Knowledge." That title is the "hard copy" version, with the online title phrased as a "question," not as a "statement." Full disclosure: If you are not a subscriber to The Wall Street Journal, you may well face a paywall when you click the link. My apologies if that is true, but even if you can't read the column yourself, you can read what I have to say about it right here. Probably, that is pretty close to the same thing.
Ip correctly points out that Artificial Intelligence (or "AI") does not end up adding to our stock of human knowledge. AI, like a "Google Search" on steroids, roams through the entirety of human knowledge already out there, and responds to our inquiries by mobilizing that knowledge, "already known," and then presenting us with a response to whatever inquiry we might have made to our AI agent, or to whatever assignment we may have given to that AI agent.
Aside from the fact that AI, currently, sometimes "hallucinates," it is true that AI tells us what we (collectively) "already know," and that it responds to our inquiries about the world by accessing knowledge already "out there." When you or I make an inquiry, or assign a task to ChatGPT, or to whatever our favorite AI agent may be, we will often have no idea or information about the topic. However, the AI response to our questions and directives, whatever they are, will be formulated from information and knowledge that is already known, and upon which the AI agent or program has been "trained."
Greg Ip's article suggests that this fact about how AI actually works will, essentially, mean that using AI will never generate any "new" idea or new "knowledge." AI doesn't "think," itself. It is just a super-effective search engine. Were we to pose a question to ourself, we would have to do the research individually, and we might come up with some completely new idea, something never ever thought about before.
Since I do think that this is a fair way to describe what "AI" does for us, if we use it, the "question" or the "statement" addressed by Ip is truly on target. We know that people are more and more relying on AI agents, not human beings, to respond to inquiries, and to formulate ideas, and to take action to achieve a particular result. And AI is pretty good at doing that, too, or so I am told, and because that's true, AI is putting a lot of people out of work - and that phenomenon is just beginning.
Also just beginning is the impact of AI on "education" and "student learning." There is no need for any student to research and think about any topic, whatsoever, when comprehensive knowledge is already available, and can be dispensed in minutes.
Ip's basic point is that AI will never add to the stock of human knowledge. All of its efforts will be based on the human knowledge to which it already has access.
Is this a good thing?
Not the way I see it!!
No comments:
Post a Comment
Thanks for your comment!