Readers of my blogs will know that I am very keen on keeping myself up to date. As a result, I have recently being undertaking some training in the digital world and found myself at a conference in Derby which covered the subject.

I was blown away by one presentation which included an exercise in creating a website from scratch using artificial intelligence or AI as it's known. In 5 minutes, it had produced a poor but usable website with a URL, a landing page, text and suitable graphics.

As a result, I decided to use two pieces of software as an experiment to see if I could produce usable blogs for my own website.What follows contains my experiences which hopefully you can learn from.

I decided to experiment by creating a blog on the subject of risk for charities. The original article had to be significantly edited as it read as though it had been produced by a machine and was extremely wooden in its style. The downside was that the whole exercise took me twice as long as it would normally have done. However, on the positive side, when producing an outline of the content it gave me one or two new ideas that I hadn't considered.

My wife who also likes to keep herself up to date had gone on her own digital marketing training course and came back with the wonderful piece of news that I could get the machine to write an article in my own style. This I thought would save me lots of time and effort. How wrong I was!

I decided I would write an article on the subject of what to consider when recruiting a CEO for a UK charity and asked it to write the article in the style of Elliot Harris. I decided to try out two pieces of software. The first was ChatGPT and what it produced was hysterical. I knew we we're on the wrong track when I looked at heading. It started “Navigating the Uncharted Waters of Charity Leadership: Selecting the Ideal Chief Executive for your UK charity”

It then got worse.

The first sentence started as follows: ”Greetings my fellow advocates of philanthropy and social change! I invite you on an intriguing journey as we delve into the intricate process of recruiting a CEO for a charitable organisation based in the United Kingdom.” I won't bore you with the rest of the two pages but it didn't get any better.

The second piece of software is called BARD and may have got its name from a character from JRR Tolkien’s The Hobbit. In complete contrast this was a mere half a page with what I can only describe as statements of the bleeding obvious.

The first two sentences will give you an idea of just how basic it was “Recruiting a chief executive for your charity is a critical decision. The right CEO can make all the difference in helping your organisation achieve its mission and goals.”

It then gave 4 tips as follows:

1. Start by defining your needs
2. Use a variety of channels to find candidates
3. Screen candidates carefully
4. Make an offer

As you can see, I didn't bother with the rest of the four tips as it wouldn't take an expert to see what comes next.

So, what have I concluded from this exercise?

1. AI is not yet in a situation where it can produce anything like the finished article.
2. However, what it can produce if you give it the right instructions is material that you can work on to produce meaningful copy.
3. You must be specific in the instructions you give it. For example, I asked it to produce an article in the style of Elliot Harris. What I didn't realise that on LinkedIn alone there are 18 pages of Elliot Harris names. What I should have said was in the style of Elliot Harris of Strategy and Governance and then I might have had far less work to do.
4. And finally, I decided after all this that the article I proposed to write was not necessary as there are far better people than me who could write this article and to have the right experience to recruit the right CEO

AI is not going to go away and we have to learn how to harness it to get the best use out of it.