An AI will replace my job sooner than yours

Man in front of NOC screen
Michael Björn

Head of Research Agenda and Quality at Consumer & IndustryLab

Our recent Ericsson Consumer & IndustryLab Creative Machines report states that as many as 1 in 5 think their own job will be replaced by a machine before they retire. Is that really a realistic figure?

As it happens, the high scenario in a recent McKinsey report is that 15 percent of the global workforce will be replaced by machines by 2030. Even though the significantly lower mid scenario is the one the authors of that report find more likely, it might be worthwhile to at least consider such a situation.


And while it might seem easy to talk about other peoples’ jobs disappearing, what if it was your job we were talking about? Or what if it was my job? Can I really be replaced?

The more I think about it, the more obvious it seems to me that I can. In fact, my job might be more suitable for an AI replacement than most other jobs.

Here’s why.

My job is very much about different types of gap analysis. I am a consumer researcher and many of my insights are generated by finding interesting deviations in behavioral and attitudinal trends, that I then conceptualize.

Until now, I have been thinking that the conceptualization part is the most important, and that the ability to combine analytical flexibility with a knack for storytelling is something a machine could never do. Not only does it take a human, it takes someone exactly like me to pull that off.
But if I am unique, maybe that uniqueness is just in the amount of self-delusion I have managed to drum up.

In reality, an AI could probably make such gap analysis more efficiently. At Ericsson Consumer & IndustryLab, we collect more than 100,000 interviews every year in what we call our Analytical Platform. Having an AI sift through several years of hundreds of thousands of interviews with millions of data points to find the most intriguing changes would be much more efficient than I would be. Indeed, we are already thinking about how to employ machine learning to help us with this.

Another kind of gap analysis is to scan all that happens in the consumer technology space and then survey consumers about their attitudes to the most nonlinear aspects of those changes. An intelligent machine would quite likely be much better and more thorough at finding the most significant gaps and designing such surveys than me. Honestly.

Then there’s the work I try to find time to do as an adjunct professor at the university. Writing papers is really very much about reading other papers in a narrowly defined area and then again doing a kind of gap analysis to find out what angles have not yet been covered. Maybe I am still better than an AI at writing the actual paper, but the problem is that I seldom get that far due to lack of time (and patience!) to scan what everyone else is writing and define an opportunity. I think an AI would beat me in getting an actual paper published already today.

But that’s not all. There is also all of the presentations and lectures I do both in business and academia. But here’s a secret – a lot of those are about reuse and repetition. Maybe that partly shows how lacking I really am in the creativity department, but I also believe that the emotional value of having a person rather than a recording or a machine is really what gives me the upper hand when being in front of an audience. Once we get used to having avatars represent us in video calls maybe that will start to change. At some point an AI could do the job without audiences noticing any difference.

But, how human am I anyway?

Cambridge philosopher Nick Bostrom postulated in his famous paper that since everything will eventually be possible to simulate in a computer, maybe that has actually already happened and we are in fact already living in a simulation. There are some famous adherents to this idea, like Elon Musk for example. However, it turns out Bostrom was probably wrong. Simulating the universe would need more atoms than the universe contains. So if something should be simulated, it needs to be smaller. Like humans, for example. What if all jobs at some point can be performed by machines instead of humans? And if so, what if that already happened and we are all AI machines even though we don’t know it?

Personally, I certainly don’t think that is the case. The point is that creative professions are not immune to AI replacement. But that does not have to be all bad news.

As we say in the Creative Machines report, those who become more skilled at operating an AI system will gain competitive advantage and increase their productivity to the extent that they can do the work of whole teams.

Imagine if I could do all the work tasks I have described above as a so-called centaur, that is, with an AI that augments my intelligence rather than competes against it. It would be extremely exhilarating to be able to generate more insights while at the same time being able to make them bolder and more impactful. And I do not think it would have to come at the expense of my colleagues losing their jobs. We have more requests than we can handle, so with some luck, the outcome would be that we would all become empowered and accomplish more than we ever dreamt of!

So maybe I actually welcome the day when my job gets replaced by an AI. My current job, that is. Then, maybe, I can do all that I only dream about today!

Like this post? Find out more insights from Ericsson Consumer & IndsutryLab’s Creative Machines report.

Michael Björn
Michael Björn is Head of Research Agenda and Quality at Consumer & IndustryLab at Ericsson ConsumerLab and has a PhD in data modeling from the University of Tsukuba in Japan.
The Ericsson blog

In a world that is increasingly complex, we are on a quest for easy. At the Ericsson blog, we provide insight, news and opinion to help make complex ideas on technology, business and innovation simple. If you want to hear from us directly, please head over to our contact page.

Contact us