The AI Article
I’m not the quickest to comment on the latest tech trends. I read. I pay attention. I experiment. I take it all in. I read the hot takes just like everyone else.
Artificial Intelligence is a hot topic, to say the least. I’ve intentionally waited a while to give my opinions some time to settle into shape. Enough time has passed, and now I’m ready to share my insights. Of course, as always, this is more about the technology leadership issues, than the technology itself. So, here we go. This is the AI article.
Large Language Models
Artificial Intelligence capabilities such as Machine Learning have been commonplace for some time. That may have been a cool blog topic 5 years ago, but not today. Our company has been putting machine learning to work for years and yours probably has been too.
The new element here is the Large Language Model (LLM). The most popular one is ChatGPT, which launched late last year. It ramped to 1 million users in just 5 days and now boasts an estimated 100 million active users. Through a simple chatbot interface, ChatGPT can generate elaborate and unique content. This is brand new. And with all new things, sometimes we react in funny ways.
first, we’ll discuss some problematic leadership responses, and then we will find our way to an appropriate position.
Don’t block it
I’m amazed at the knee-jerk response of tech leaders to block what they don’t understand or what they cannot completely control. It always comes across as weak leadership to me. I also find it ironic, since we are supposed to be futuristic people.
Firstly, blocking it rarely works. Sure, you can block it on the corporate firewall, but people can just use it on their phones. All emergent technologies have a mixed bag of benefits and risks. It doesn’t make sense to mitigate some of the risks by depriving yourself of all the benefits. It makes more sense to set a guideline and educate people instead.
I’m old enough to have lived through the BlackBerry to iPhone/Android transition. When iPhone/Android first came out, it wasn’t nearly as secure or corporate-friendly as the BlackBerry. Not by a long shot. It was, however, very appealing to end-user productivity, and that was enough to easily win the war.
Folks, let’s not repeat history. I remember the tech leaders that lovingly clung to their BlackBerrys, espousing the risks of unmanaged iPhone/Android phones. I don’t think it worked out too well for them in the end.
Don’t anthropomorphize it
AI isn’t a person. It’s a tool. It’s a tool built on an expansive base of human knowledge. If it were a person, it would be the most unoriginal, boring person on the planet, because it’s designed to be average. Real people aren’t average. We’re all unique and quirky in our own way. AI is a very powerful tool, but it’s a long way from personhood. It can generate content, but it is not creative. Know the difference.
I realize that AI pushes this boundary. The most impressive example I’ve witnessed was AI finishing a symphony from a dead composer. The AI used the composer’s previous works as training data and picked up where he left off to finish the piece. I listened to the symphony played in Carnegie Hall by Julliard students. There’s nothing about it that seemed computer-generated. I’ll argue that the dead composer and the student musicians were the creative ones, not the machine.
Now that we covered a few of the positions to avoid, let’s embrace the path forward.
AI is a tool
I’m writing this article with the assistance of a spell checker and a grammar checker. I’m not all that good at spelling and grammar without the computer. Should I add a disclaimer to my article that it was written with the assistance of spell check? No, I don’t think so. I’m still the author. Sure, the computer helped, but not as a co-creator.
When I was in grade school, I was terrible at math. I knew how to do it but was painfully slow. When I got into junior high school, they let me use a calculator for rote math, and then I learned abstract math. I was fantastic at that. I doubled up on courses and eventually aced AP Calculus my senior year.
AI is a new spell checker. It’s a new calculator. It’s a productivity enhancer. It’s not cheating. It won’t eliminate jobs. Well, it won’t eliminate the jobs that are worth keeping. We don’t have many workers building typewriters anymore either.
LLM AI is a major leap beyond calculators and spell checkers but is fundamentally still a tool. Tools tend to make humans more productive and do not fundamentally replace people outright.
AI has biases
Much has been written about the biases and ethics of LLMs such as ChatGPT. If there’s something that needs a disclaimer, this might be it, but for me, this is also expected. Some may think biases need to be added for ethical AI. Others think the biases programmed in, are the ethical problem.
Fortunately or unfortunately, the fact is, that all of our tools have biases in them. They are all made by humans that have some sort of agenda. This also isn’t new.
Wikipedia articles are biased, but I’m still grateful for a free and easy-to-use online encyclopedia. Social media content moderation is biased. While it would be cool if those biases decided to boost my ZoL posts for a day so I could enjoy 15 minutes of fame, that’s not the way the programmers designed it.
Lastly, but most notably, the news is biased. We all know that, but we’d still rather be informed than live under a rock.
So yes, AI can act like a human and could manipulate you according to its programmed biases. That’s not great. It’s a risk. Here’s the question to answer: Is it so bad that we should shut it off completely? As a technologist, I think we should proceed with our eyes wide open.
AI could destroy the world
It’s a worst-case scenario, but I’ve seen enough science fiction to keep an eye out for the uprising of our new AI overlords.
A few weeks ago, a group of notable AI leaders authored a warning about the existential dangers of AI to humanity, recommending a pause of at least 6 months. Among the signatories was Elon Musk, who has been vocal about sounding the alarm on AI for some time now. He’s one of the founders of OpenAI, the maker of ChatGPT, and he’s also developing a brain-computer interface called Neuralink, so there’s that.
The warning is salient and worth reading. However, I don’t think any development will stop for a millisecond. There’s too much to be gained.
The technology leadership path forward
Tech leaders should simultaneously embrace AI while working to understand and mitigate the risks. On one extreme, you can try to resist it now and stick your head in the sand. On the other extreme, you could embrace it with wild abandon, driving us straight into the apocalypse.
The best tech leaders are on the leading edge, carefully experimenting with eyes wide open. As always, the future is both exciting and dangerous. Proceed accordingly. We are the pioneers. That’s our job.
Podcast: Play in new window | Download