I was thinking back to when I joined the Jaycees a couple of decades ago. I have long since aged out. After one meeting, a younger member and I went out to brainstorm some business ideas. He was surprised to see that I had most of my important numbers memorized and was manually punching them into my flip phone instead of storing them as contacts. This was around 2005.

Fast forward to now. I keep seeing articles about how conversations worth having are becoming a lost art, something mostly associated with older generations. It got me wondering: Are technology and specifically AI making us dumber? Is our fate already a foregone conclusion?

Last night, I was reading Understanding Deep Learning by Simon J.D. Prince (MIT Press, 2023), which I will be reading for many nights to come, and it sparked a thought. While I’m not overly worried about relying less on rote memorization, what does concern me is the potential erosion of creative and critical thinking in an AI-driven world. Deep learning itself is built on neural networks, systems about which there is still far more we don’t know than we truly understand.

But here’s the flip side. Maybe the very concepts behind how deep learning models are designed could actually help revive a golden age of thinking, bringing us back to a sharper, more thoughtful way of solving problems. Maybe studying how neural networks process information could remind us how to think more like the great minds of the past, e.g., Antisthenes, Galileo, Jefferson, who were masters of careful, iterative reasoning.

Here’s a specific question for the forum that’s been rattling around in my head:
What can the concept of backpropagation in deep learning teach us about how we process information and make decisions, especially when drafting or negotiating something as complex as legal contracts?

I am curious what others think.

@Guardian Great question! Unfortunately for me I know next to nothing about machine learning and AI and have little interest in it. But I think there are others here, like Don Mansfield, who might engage with you. Thanks for posting.

I agree that's a great question. I don't have a thoughtful answer.

As for the concern that AI is making us dumber, I'm not so sure. I observed my son using ChatGPT to write a paper for his high school class. There was definitely creative thinking as he continuously revised his prompts and critical thinking as he reviewed each new product to see if it was any good. I certainly didn't think he was cheating.

1 hour ago, Don Mansfield said:

I agree that's a great question. I don't have a thoughtful answer.

As for the concern that AI is making us dumber, I'm not so sure. I observed my son using ChatGPT to write a paper for his high school class. There was definitely creative thinking as he continuously revised his prompts and critical thinking as he reviewed each new product to see if it was any good. I certainly didn't think he was cheating.

He was not cheating, @Don Mansfield, and I'm sure many teachers prefer it to the alternative. As you implied, it all depends on how one uses it. Would anyone say that using Westlaw or LexisNexis and their advanced search engines is "cheating," that lawyers should instead be poring over decrepit bound volumes in the stacks? I feel like ChatGPT is making me smarter, or at the very least, more efficient and knowledgeable. As I’m reading a complex book on technology and thinking about the intricacies of federal contracting in the back of my mind, I can pause and pivot, taking a call from my brother and thoughtfully answering his question about the effectiveness of barometric fans in an attic and their ideal placement. I hang up the phone and go right back to my book. Moreover, how else could someone with a background in law and ethics possibly sit down and inhale 500 technically-dense pages from MIT before interlibrary loan demands its return?”

Anything I don’t understand or need explained more clearly, ChatGPT is great at simplifying. In fact, I recently read an article that said ChatGPT excels at breaking down complex tasks and offering real-world analogies. My experience has been that it does. While I don’t grasp every detailed exhibit in this compendium on deep learning, I wake up feeling more confident in my understanding of the concepts and find myself asking uniquely thoughtful questions as I water my parched lawn. As someone who has always considered himself more right-brained, this is the most effective way for me to learn about the science of artificial neural networks. Sometimes, you have to work through something with these chatbots a dozen or more times before you get what you want, but the technology is surprisingly advanced. There are legitimate concerns that these tools are hullicinating their way through the preparation of legal briefs. But for anyone who is censured by a judge or the like, they have only brought it upon themselves. As an academic, you read one book, then another, and another. You then sit down and write. You don’t copy from a single book. That’s plagiarism.

Law is extremely nuanced and purposely so. For one, its methods protect a class that has spent a decade investing in their expertise, time and money that could otherwise have been directed toward a passive income-generating instrument with compounding returns. The same holds true in the medical field, at least insofar as its use of Latin and Greek to describe concepts that are otherwise manageable and accessible.

2 hours ago, Vern Edwards said:

@Guardian Great question! Unfortunately for me I know next to nothing about machine learning and AI and have little interest in it. But I think there are others here, like Don Mansfield, who might engage with you. Thanks for posting.

Thank you, Vern. I always appreciate you.

I imagine, as always, there will be folks who use AI to copy and paste without perusing the contents and lose their ability to think and know how to apply the vast information that is available through AI tools. On the other hand, the paradigm of learning and producing work is already shifting. The competitive edge will come from learning how to formulate optimal prompts that extract precise and accurate results from AI and knowing how to curate those information and orchestrate a case that strengthens your work. Much like a maestro, who is not required to know how to play all the instruments in the orchestra, but knows how to command and arrange the instruments together and conduct each sections harmoniously to convey the best interpretation/representation of the score.

Relatable, too simple or not the discussion raised this thought. A person now uses a calculator to do math or uses a digital watch or clock to tell the time - cheating, dumber? I think not, yet the experience of knowing what I will call long hand math or using a analog watch in my view adds substance to the use of the former.

This type question always gets asked by every generation - how will “X” affect the way we think and act? For those young, inquisitive, and eager, machine learning and AI is just a logical evolution in technology. Adoption is second nature. Their answer likely will be something like it makes us all better without dwelling into anything deeper like renaissance thinking.