It was one year ago this month that I had the crazy idea to learn to code. Actually, my goal was to discover a way for artificial intelligence to write my weekly column and books.

I dreamt about the hours it would save me and how cool it would be to blend natural and artificial intelligence. I had read an article about AI writing poetry and a Harry Potter chapter, so I wondered, why can’t I train AI to write like me, then provide a topic and minutes later have my column? (Spoiler alert: this week’s column is written by me, as are all the others!)

Well, it turns out that it’s not that easy, and from the outset, my lack of technical background made my foray into the world of coding all the more challenging. Googling “How to write with artificial intelligence” yielded pages and pages of links outlining how to go about it, why it’s brilliant, but ultimately, why it can’t happen — yet.

Still, being the person that I am, I thought I’d give it a good olé college try. It couldn’t be that hard to train a machine learning algorithm to write like me, surely. One of the hardest parts, according to the advice sites, was having enough training content, but having written thousands of columns and a handful of books, I was sure I had enough.

Unfortunately, however, I hit a wall almost before I had begun. As a start, I downloaded and installed TensorFlow, an open-source software library for dataflow programming produced by Google. Next, I copied in some code to load the data.

The problem was, I had no idea if it was right or not. That would’ve been the perfect time to throw in the towel and forget about my ambitious idea, but then it dawned on me that I should hire a coach.

With some expert help, I reinstalled TensorFlow, correctly this time, and fed it with everything I’ve ever written — every book, article, and paragraph. But it wasn’t enough. The output I received in the sample paragraph was nothing more than a confusing string of random words.

I was disappointed, but it seemed like a simple enough obstacle to overcome. Undeterred, I uploaded dozens of my favourite books: all the writings of Malcolm Gladwell, Walter Isaacson and JR Moehringer, as well as titles like “Shoe Dog”, “Grit”, “The Decisive Moment and Thinking, Fast & Slow”. After all, these great works, blended with my own writing, would make me even better.

A win-win, so I thought.

So, my coach and I ran the code to load the training data, then ran another set of code to spit out a column. But, as the words appeared on the screen, I shook my head in disbelief. The results were nonsensical.

Then, in the midst of this experiment, my coding coach told me something I didn’t know, and it was something that changed my life: there is a free library of algorithms, and everything in it is open source. I had assumed that every algorithm was bespoke and that to be able to have artificial intelligence, you had to be able to write your own, but I couldn’t have been more wrong.

When my coach told me this, the lightbulb switched on in my mind, and the question shifted from, how can AI write my column for me? To, how can we use AI to make leaders better?

Since my mid-20s I’ve woken up every day impassioned to help leaders grow, and the realisation that we can use AI to do just that has fuelled my passion even further. The coming transformation is machines learning, humans leading.

That is a pretty big gap between what I wanted it to do and what it can do. But that weekend transformed my life. What was once untouchable was now approachable.

I realised that nearly everyone in the world can use machine learning. How will it transform your future?

— Tommy Weir is CEO of EMLC Leadership Ai Lab and author of “Leadership Dubai Style”. Contact him at