AI is not a good leadership trainer, not because it never gives good information, but because the pool of information is muddied, and you don’t know what grime you will get mixed in.
Should you use AI to train in leadership?
Generally, no.
It’s not that it never gives good answers or can be right at times. But with the way AI works and the amount of not-so-great content out there, there are high chances that what it teaches you isn’t accurate.
I’ve personally had it share with me a popular leadership model to help prove one of the points I was researching, then in another search tell me how that popular model wasn’t backed by empirical evidence.
This is just a sample of the issues that can come up. Why does this happen?
Let’s take a deeper look.
AI amplifies what it sees – and much of the source material out there isn’t good
AI is pattern recognition. It pulls from the sources out there (whether articles, videos, social media, etc.), and it puts it together and pulls from it.
The problem with leadership is that it is one of the most written and talked about subjects, and much of the content out there is wrong (or at least superficial or misguided).
There’s a few reasons for this.
One, leadership is a buzzwordand is used as a catch-all phrase for many topics: entrepreneurship, confidence, personal and character development, and more.
Then there is the Dunning Kruger effect that adds on to this. People learn a little bit, or they get into a position, and they think they know leadership. So they write or post about it.
And many of the articles written on leadership aren’t made by leadership “experts”, but by marketers, for example. They want to get SEO status, so they write on the topic, and to do so, they pull from what everyone else is saying, so they often repeat things that aren’t accurate just because it’s popular, so mistruth keeps getting printed.
Much content out there on leadership is repeated because its popular and feels good and trendy, not because it’s true or accurate.
For social media, anyone with a camera or keyboard can post about leadership confidently, whether they are correct or not. And with leadership being a buzzword, meaning almost whatever you want it to be, then the concept of leadership and what makes a leader effective gets even more muddled.
Even leadership research has its issues
Leadership research has its own issues. Even within the research realm, Leadership is given many definitions. In fact, one researcher (Bennis and Townsend, 1995) reported an estimation at least 650 definitions of leadership in the literature – and that was in the 90’s!
Not only that, but many studies today are cross-sectional, one-time surveys that may show correlation, but you can’t find cause and effect with them, even if they try to. Then questions are made to fit the surveys versus finding the research method to fit the question needed answering.
And then construct proliferation (creating more and more leadership types and styles and approaches or formulas) makes it worse (the same dynamic shows up across I-O psychology as well).
Researchers have the incentives to create new constructs, whether they are needed or not, because it looks good on them when they do.
And then (and my personal opinion), it seems some are adamant about certain “types” of leadership, so they do everything they can to show what they think is right is true (a typical human thing, but still hurts the research when that’s true).
In summary, AI is drinking from a dirty well, then serving it back with confidence and fluency.
AI likes to validate
On top of that, AI likes to validate and hallucinate. Now, you can tell it to argue against you and tell you if you are wrong or rate the claims it makes.
However (and each agent is different at what level it does this), it likes to tell you how wonderful you are and tell you why your ideas are great.
If you say something completely wrong about leadership in a question or statement, it’s likely to agree with you and show you why you are right, even if it invents it along the way.
That, on top of much of the confusion and misinformation about leadership already out there, it can lead you in the wrong way.
Where AI Is Useful
Now, you could say that this applies to all content, and it does, to an extent. But leadership is also a highly written about topic and one of the topics that has so many different opinions and confusions about it.
When I’ve used it for research on leadership, and I compare it to what’s accurate, sometimes it does a great job.
And sometimes it doesn’t.
I’ve had it be spot on, and I’ve had it say things that could even be harmful if you practiced it as a leader.
This doesn’t mean you can’t and shouldn’t use it as you learn leadership. But you have to be careful. You need to start with a good basis of what leadership is.
While there is confusion about leadership and how to define it, many credible frameworks consist of the same things, similar to how Peter Northouse defines it:
“the process of guiding people toward the accomplishment of a goal through influence” (paraphrased).
It’s important to recognize not everything is leadership. Not everything is good. Not everything makes you effective.
Just because it feels good or sounds right or is trendy or is popular or seems great doesn’t mean it is.
If you want to use AI for leadership learning and training, make sure you have a solid foundation first. Don’t use it for your foundation.
But once you have the foundation, you can use it to explore and research and build on it. Just make sure to validate it with the foundations you’ve learned (and the nomal verifications with AI and sources).
And if you want a solid foundation, take our comprehensive, no-cost leadership training here.
