On Ai

AI was the subject of a recent article, where it was praised as a positive influence in the lives of not only the kids but also the teachers. We are bombarded with programs and new articles about how AI will take over the world and kill us all, but very few of us have taken the time to understand its capabilities and limitations. I hope to change that here.

Before we begin, I would like you to consider something few of us think about. How can we carry on a complex conversation with friends and family for hours at a time without missing a beat? We never stay focused on a single topic; we jump around, get distracted, get interrupted by people trying to discuss something else, and argue over the meaning of a single word before moving on. How does your brain accomplish this? How does something someone told you months ago, or something you read years ago, help you develop a better explanation of a topic? Most of the time, they’re not even related thoughts.

How many times have we driven down the road and had someone ask, “What are you thinking about?” only to realize the insane track your brain took to get there? Any attempt to explain the transition would only make you look insane. Consider how often you get distracted, thinking about something entirely separate from what you’re doing. Our brains are pure chaos, yet this chaos is where our creativity comes from. We’ll get back to this point shortly, but first, let’s consider how AI operates and why it isn’t what is claimed.

First, we need to consider that it learns from books, research papers, legal decisions, photos, images, designs, blogs, Twitter, videos, and more. How many books are written with opposing views? How many YouTube videos have you watched that claim the opposite of another? How many politicians and attorneys argue the opposite of each other as an occupation? So, it is safe to say that someone has to direct these databases to store information on only one viewpoint or another. When you listen to an answer from an AI, you are being presented with an opinion, dressed up as fact, from someone, somewhere, at some time, and your ability to consider the opposing viewpoint is removed. Worse, because you use the AI to answer basic questions you deal with daily, you’ve assigned it a characteristic you’re not even aware of.

Let’s consider the process: when you ask AI a question, it has to store that question in memory, then goes out and searches its biased database for possible answers, which it stole from other people’s websites, articles, and books, and then presents them to you as if they were its own original thought. The same thing would have happened if you had read the five websites it retrieved the information from, except that while you were reading, your brain would have gotten stronger, and you could cross-reference and determine which version was the better answer. But when you used AI, you lost all of that. This isn’t even the truly insane part of what is going on. What happens when the answer comes back, and you decide to push back or refine your question, so you ask it something else?

At this point, the AI must go back to your original question, read it, then read its response, then read your next question, and perform a new search with refined search terms. It then repeats the process to provide an answer, appearing to think but, in fact, only performing the functions of a search engine with a single answer. If you still don’t get it, push back a third and fourth time and extend the conversation to an hour or more. Each question requires the AI to go back to the very beginning of the conversation and to read question, answer, question, answer, all the way back to your current position. This isn’t how your brain works; in fact, no one has a clear idea of how your brain handles a similar conversation, especially while using so little energy.

Now, regarding the concept of Generative AI, which was alluded to in the previous article, we are nowhere close to achieving this level of AI, as it would require the system to work like our brains and combine unrelated concepts to create new things. It would have to run in chaos mode and try to be creative. What we currently call AI is really just a search engine combined with a speech engine to form sentences we understand. Its answers, while appearing creative, are really just elaborate math solutions that try to find answers and use associations to form, hopefully, a coherent response.

What scares me more than teachers not understanding the details of AI is that they are letting our kids use it in school. If someone had to research a topic, read multiple articles, and then offer their opinion on it, how could that be considered accomplished if they simply asked someone else to do it for them? That used to be called cheating.

AI is an awesome tool; however, at the end of the day, it is still a tool. I don’t believe our kids should be allowed to have access to a phone, computer, or tablet while at school unless doing research. Suggesting they are better off using AI to finish their homework faster is something altogether askew with education. It also offers further proof that what they call education should be renamed “Schooling,” as John Gatto called it years ago. As he remarked in “Dumbing Us Down,” something we all should read, “It is time that we squarely face the fact that institutional schoolteaching is destructive to children.” Suggesting that using AI is a positive for schooling should be a red enough flag to have you searching for other options.