Roger Bootle
The whole question of AI interviews is a bit like warfare really. Every technological advance on the side of offence is met by technological advance on the side of defence. Similarly, there are now algorithms that prepare candidates how to approach AI interviews.
I suppose in the early stages of the interview process there is something to be said for going down this route, although I must say it is completely against my own instinct. When I was actively running Capital Economics a lot of my time was taken up interviewing people which was one of the most important jobs that a CEO like me could do in a small firm to make sure that the people were really good. In those days, I was aware that what I was looking for wasn’t necessarily straightforward so I would have loathed giving up that initial sifting to AI. For a start you get a bunch of paper CVs in and I could tell very quickly whether a person was plausible or not. I could sift very quickly. When you come to the next stage, and you have got rid of all the implausible applications, you certainly wouldn’t want an AI algorithm at that stage.
Really it goes back to this whole question of being human: although a lot of people will resist this, particularly in small organisations, hirers will be very motivated by whether they think they can get on with this or that person. This is especially important in a small outfit, though I accept things might be different for a bigger company. If you are working at close quarters in a small company and you really don’t like a person, that’s an important negative. How is your algorithm going to pick that up?
It’s quite possible therefore that AI interviews will never really take hold for smaller companies. Whereas with those big companies, where there are hundreds if not thousands of applicants, that basic stage could be very time-consuming so employers might find it efficient to get AI to tackle that. You might miss the occasional good person and let through the occasional duffer but you can sort that out later. In a small company, this sort of thing matters so much more.
For reasons I won’t go into, I happen to own a pet shop and dog grooming salon. It’s a niche business, and so it’s very difficult to recruit staff. We happened to need a new manager, and we advertised for the role. The manager is critical – a good manager will take the burden off me and make the business thrive. It was 2020, and I happened to be in France and I conducted the interview by Zoom. I rejected the candidate who everybody else thought was the best of the bunch, and hired someone over Zoom who turned out to be a disaster. Zoom obviously isn’t AI, but it is similar in that I didn’t have the sort of human contact that I would normally have in an interview process of meeting someone in person.
In a similar vein, there is also a fair amount in my book The AI Economy about education. Some AI enthusiasts say that there aren’t going to be any teachers any more because people can learn remotely from various programmes and so forth. I strongly reject this idea. I would recommend Sir Anthony Seldon’s book The Fourth Education Revolution. According to Seldon, there is scope to use AI a lot in the education process, but the system of the teacher standing up in front of a class of sometimes hundreds of people and the students taking notes is ludicrously antiquated.
Instead, I suspect education will proceed along the lines of the tutorial system whereby we will have more one-on-one sessions which are about discussion and interaction, in addition to seminars where you have got a small number of students discussing and interacting. Under that system the ratio of teachers to pupils or students in aggregate may not change that much but the ratio in individual teaching sessions will change dramatically.
AI won’t change our lives anything like as much as the enthusiasts claim because we’re human beings and we will always crave some degree of human contact across every area of our lives.
The AI Economy by Roger Bootle is published by Hachette UK and priced £20
