AI and Leadership

I was in a conversation about AI and Leadership recently and a number of themes were common while some were less so.

We were discussing someone who was trying to position themselves as a thought leader by regularly posting articles that showed all of the hallmarks of having been written by AI.

Using AI is fine. I have no problem with people who use AI. That's not my problem. My problem is when someone posts, publishes, disseminates, etc. a piece of content and calls it their own when they did not, in fact, craft it. My problem is when the responsibility for human judgment is abdicated in preference for expedience and speed. My problem is when fear and a desire for comfort edges out creativity and risk.

In my mind, these are not the traits of a leader.

I'm going to get a little technical here. AI (Artificial Intelligence) is often used to refer to a variety of technologies, including generative LLMs (Large Language Models).

So it's clear, I'm using AI and LLMs interchangeably. In reality, they are technically different (generative LLMs are a subset of AI).

LLMs are trained on enormous amounts of text and other data, producing very large and complex models. In broad terms, they generate responses by estimating what text is most likely to come next based on what came before. For example, if I provide, "A giraffe is," the words that might naturally follow include "tall," "a mammal," or "four-legged." The model has learned patterns that help it estimate which continuations are more or less likely. Add in some controlled variability (temperature) along with the prompt, context, instructions, and other settings, and the LLM generates a response.

In essence, LLM technology is very complicated auto-complete.

Those models are built using a process called "training." To train a model, a computer program is provided with a massive amount of data from which patterns, relationships, and probabilities are identified and represented within the model.

Please allow me to point out a key fact here: when a model is trained, words and thoughts that other people have already spoken are fed into a program for analysis. What comes out is generated from patterns learned from human expression. That does not make it meaningless, but it does mean leaders should be honest about where the words came from and what role their own judgment played.

If you use AI to generate thoughts and words, you are dependent on patterns learned from what people have already thought or said.

An LLM builds patterns based on the data used during training. As a result, that which is generated by an LLM is built using patterns found in the thoughts and words of other people.

Just now, as I was typing this, I accidentally included a typo. I wrote, "generated by and LLM" when, in fact, I meant to say, "generated by an LLM." All of the words were spelled correctly; however, because of the many, many, many times "an LLM" were observed during training compared to the very, very, very few times "and LLM" were observed, the editor reasonably concluded that while I wrote "and LLM," I most likely meant, "an LLM" and provided a little blue squiggle underline beneath my word choice.

The technology is far more complicated than that; however, it serves to highlight my point: it's far more statistically likely that one word appears than another word, so the LLM drew the conclusion that I messed up. It was correct... this time.

Using AI is fine and often very helpful; however, presenting the work of an AI as your own without citing that it was created using AI may be seen as dubious or even dishonest, particularly from an academic perspective.

Consider a student writing a paper. The student finds a great research paper, essay, or book someone else wrote. They copy the words into their own work and submit it under their own name. Without citing the sources, this is called plagiarism. AI changes the mechanics, but it does not erase the ethical concern. When someone presents AI-generated words as though they reflect their own thinking, judgment, and craft, they are asking the reader to believe something that is not fully true.

Using exclusively AI-generated content and presenting it as your own does not make you a leader. In fact, in my humble opinion, it does the opposite.

When content — essays, posts, articles, books, etc. — are presented as one's original work, then truth becomes hidden, muddled, and obfuscated. This is not leadership. Twisting the truth to make ourselves look more original, insightful, or authoritative than we really are is not leadership.

When exclusively AI-generated content is presented as original leadership thought, the issue is not merely whether the phrasing is new. The issue is that the leader's own lived judgment, risk, and accountability are missing.

That is not leading, it is following.

So, why would someone do this? What's their motivation?

Those are good questions. I don't know the answers. I can guess, but I can't know.

I want to make it clear that I'm not saying that AI or LLMs are bad or unhelpful; I'm not saying that using an AI is dishonest. I used AI to verify my grammar and consistency in this writing. The difference here is that I'm presenting my own thoughts with my own words. I used my own judgment and reasoning to determine what I did or didn't want to say. I didn't ask a computer to think for me.

I'm also not saying that interacting with an AI as a sounding board or as an editor is bad. Many of my writings came from me starting a chat with an LLM by asking something like, "I'm exploring the idea of [blank]. Will you please help me pressure-test it?" or "Are there any instances where [blank] happened? Could you help me research that, please?"

For me, the leadership test is this:

• Did AI help me say what I believe more clearly or did I use it to avoid doing the work of believing, deciding, and saying something myself?

For me, leadership is about going places. It's about a journey. It's about struggling and difficulty and growth. It's about overcoming one's fears and doing something interesting or new or creative, despite the risk — no, because of the risk — knowing that failure and humiliation are absolutely possible.

If you're going somewhere others have already gone, you're following (as in the expression, "following someone else's footsteps"), not leading.

I don't want to go there. I've already been there. I've got the tee-shirt and the ticket stub.

I want to learn and grow and experience. I want to live. I grow as a person and we grow as a society when we expand our horizons and ponder new ideas.

Similarly, I want you to learn and grow and experience. I want you, dear reader, to live. I want you to grow as a person and expand your horizons and ponder new ideas.

I don't know about you, but personally, I'm not particularly interested in being a follower. I don't want you to be a follower, either.

Leadership is not a zero-sum game. Leadership doesn't require a fancy title or special role or elevated authority. What it does require is treating others with honesty and respect and dignity. This includes the people whose work is being cited, sampled, and used as a basis for new work and the people you're leading.