The recent obituaries for Charles Handy have been suitably kind to a wise and insightful being. A wonderful raconteur capable of creating simple images to communicate complex ideas, such as the Shamrock Organisation, what I found particularly interesting were the replies to the obituaries in the papers. It seems that everyone who met Charles Handy has a story to tell about him. So here is mine!

I was speaking at a Conference around 1990. After my talk I was introduced to him. He had caught the end of my talk. My theme was why ROI in IT systems was higher when focussed on organisational effectiveness than efficiency. He wanted to know the source of the research and was pleased when I told him MIT. At that time, I hadn’t discovered his thinking around the “curse of efficiency”. At lunch, another delegate tried to argue with him that efficiency drives productivity, and in terms of economic growth, only productivity delivered growth in the longer term.

Somehow, we got around to productivity of professional work. He floated the question of how you measure the productivity of a symphony orchestra. Playing Beethoven’s 9th in 30 minutes is not a doubling of productivity. Playing with a small orchestra and chorus may work musically but is a very different experience to the wall of sound of the choral movement. I can’t find anywhere in his writings that he developed that line of thinking.eHe

So, how do you get from there to AI? Bear with me. I have been increasingly feeling uneasy with the hype around AI. While the technology advances, in my opinion, the gap between technological potential and commercial returns seems more distant than when the dotcom bubble burst. There are some who argue that the returns will be this year and others who think it’s a decade or more away. Thinking back to that first meeting, my talk aimed to show that the major investments needed to turn IT into economic value added were actually in governance, skill developments and infrastructure,

I have been following the posts of a number of former colleagues and associates around their experience of using Current AI systems. One word has come up on a number of occasions that I hadn’t expected, “enjoy”. For example “ I haven’t enjoyed coding so much in years”. Now enjoy was very much a Handy word. When told how much he earned by an individual at a think tank meeting years later, his response was “but do you enjoy the work?”

The source of my unease became apparent, when I made the link to music.

My son is a violinist. When he was about 10 he was loaned a much higher quality instrument than he had started on. To a tone-deaf person like me, his progress seemed to become more rapid. He told me that he was really enjoying having the new instrument. So, we have a chicken and egg problem. Was he progressing more rapidly because he was enjoying it more, or was he enjoying it more because of the progress he was making?

I tend to think that he had reached a level of skill where a better instrument was valuable.

If you heard me playing a Stradivarius, and Joshua Bell a £100 violin, I promise you he would still be better. If we gave everyone learning the violin a Stradivarius or equivalent, would they all play like a maestro?

Reading the posts on personal experience of AI, what clicked for me was that I became more impressed by the expertise (many over 30 years) of the writers more than the technology per se. Rather than (Artificial) AI, it made more sense to me as (Augmented) AI.

So my hypothesis is that current AI systems will increase the quality of the best performers more than the rest. I’m trying to think of AI in professional work as an instrument rather than a machine. If you don’t have the opportunity to invest the time in the basics of the instrument, how can you build the expertise to justify the better instrument?

So, I was interested when I came across a paper, quoted here: “AI-supported research discovers 44% more new materials than conventional teams”

It’s only one example, but it illustrates my concern. This is a good study in the use of AI in materials research in a company. It shows an increase in output, but also a widening gap between different research teams performance.

So, try this as an example. You are the Head of HR in a professional services organisation. How do you feel about your star performers widening the gap between them and the rest? Maybe you can let a few more go! But there is another issue here. The study indicates that some of the most productive staff felt less job satisfaction. How do you handle a situation where your star performers enjoy the work less?

I’m ploughing through other papers which hint at potentially similar outcomes in different domains. I’m not sure where my thinking will go next.

So, when will the returns on AI justify the investment, 12 months or 10 years? I think that there are returns to be made in the short term, but full benefits will probably come 10 years after you start. So, if you wait 5 years, in a competitive world you may never catch up.

My advice, for what it’s worth, is that if you are wrestling with some complex challenges, think like Charles Handy would. RIP