I have been listening to a lot of podcasts about the future of work, which is what you do when when avoiding the fact that the present state of the world really has enough to worry about.
The message is remarkably consistent. AI will transform everything. Jobs will disappear. New jobs will emerge. We must adapt, retrain and become AI fluent, preferably within the next ten minutes.
The statistics certainly sound convincing. And also contradictory.
Harvard Business Review reported that expectations for AI-driven growth remain extremely high, despite Gartner finding that only one in fifty AI investments delivers transformational value. McKinsey estimate that up to 30% of current working hours in Europe and the United States could be automated by 2030. It expects millions of people to change occupations, with growing demand for both technological and social skills.
And then, the International Labour Organization says one in four jobs has some exposure to generative AI, although it believes transformation is more likely than complete elimination.
I found all this quite persuasive. Podcasts, articles, interviews. I was starting to consider myself to be the guru on the future of work.
Then I had dinner with someone whose opinion I hold in very high regard. This may be because of his formidable intelligence. It may also be because he is my son. I accept that the assessment process was not entirely independent.
I explained what the experts were predicting. He listened and asked, “Haven’t people always predicted this sort of thing, and wasn’t it always wrong?” This was irritating. You spend years encouraging your children to think independently, only for them to start doing it in your presence. I came home and investigated.
A brief history of the future
In 1956, Richard Nixon predicted that automation would soon bring Americans a four-day working week and a fuller family life. Seventy years later, Friday afternoon meetings are still a thing.
In 1965, Time reported that computers might eventually create a twenty-hour working week and a mass leisure class. Some thinkers imagined a future in which only two per cent of the population would need to work.
Instead, we invented the smartphone so that work could find us while we were buying milk, out for dinner, or just out of surgery.
In 1975, BusinessWeek predicted that most office records would be electronic by 1990.
The electronic records did arrive. But paper refused to leave. Computers made it easier to print documents, so we printed more of them. Anyone who has ever stood beside a printer insisting there is a paper jam when no paper is visibly jammed anywhere will understand that technology does not always remove a problem. Sometimes it gives the problem a small screen and an attitude.
In 1984, one prediction suggested that good secretaries would remain in demand until 1990, when voice recognition would be perfected. Secretaries did not disappear in 1990. Voice recognition was not perfected. This week, my phone responded to the simple request to find a restaurant near Winterthur, by offering directions to a carpet shop in Austria.
In 1995, the year I gave birth to the person who sparked me to write this newsletter, Jeremy Rifkin published The End of Work, warning that technology was taking us towards a near-workerless economy. Work did not end. It acquired Wi-Fi and followed us home.
In 2005, a future-of-work organisation predicted that one third of Americans would be working away from the office by 2010. The technology was ready. Management was not. Many managers remained convinced that an employee who could not be physically observed might be at home enjoying themselves. It took a global pandemic to achieve what broadband alone could not.
Then, I went on to learn, in 2013, Oxford researchers estimated that 47 per cent of American employment was at high risk of computerisation. This was rapidly translated into “half of all jobs will disappear.” Yet the researchers explicitly said they were not predicting how many jobs would actually be automated or exactly when. A measure of possibility had become a redundancy notice.
They were wrong, but not completely
Computers did transform offices. Clerical jobs declined. Remote working spread. Machines took over thousands of tasks that we used to do with humans. The forecasters often saw the direction correctly. What they got wrong was the timing, the scale and the human response.
They could predict what technology might do. They were less successful at predicting management, regulation, economics, office politics and our astonishing ability to turn every labour-saving device into additional work.
An HBR study of 485 media articles and 570 people found that the predictions are also related to who we are. Technology specialists tended to imagine optimistic futures, economists were more sceptical, and writers and journalists expected doomsday. Apparently, it is sometimes easier to predict the prediction than the future.
What if this future is wrong too?
AI is already changing work. Pretending otherwise would be a nice thought, but ridiculously foolish. Besides, the jobless future is only one possibility.
We may end up with roughly the same number of jobs, but with the easier tasks removed and the workload quietly doubled.
One thing that really makes me think is that AI may not eliminate entire professions. Instead, it might eliminate the junior work through which people learn them. The real casualty could be the bottom three rungs of the career ladder. But if that happens, how do we start at the 4th rung? And surely, experience has to become more valuable. If a machine can produce an answer in seconds, the scarce skill is knowing whether that answer is useful, misleading or complete nonsense.
And my own vested interest in the the value of the older worker makes me think that surely this should favour older workers with judgement and context, assuming organisations do not remove them from the room for failing to look sufficiently “future-facing.”
There’s also the opportunity that AI may also make it easier for individuals to build small businesses and portfolio careers. I’ve already seen instances wher one person is able to do work that once required a team. But it is not easy.That person may then discover they have become the strategist, researcher, administrator, marketing department, technology support desk and the person who knows where the passwords are kept.
The biggest question may not be how many jobs survive. It may be who owns the technology, who receives the productivity gains and who remains accountable when the machine gets something wrong. Those are human choices, not technological inevitabilities.
So I will keep listening to predictions about the future of work. I will simply treat them as possibilities rather than appointments. There is, however, one important difference this time. In the past, humans predicted what machines would do to human work. Now machines are helping us predict what machines will do to human work.
AI can, of course, be notoriously unreliable. But even on a bad day, I suspect it has a better chance than consulting a crystal ball or asking a fortune-teller to read the tea leaves.
Perhaps Winston Churchill had the safest approach:
“I always avoid prophesying beforehand, because it is a much better policy to prophesy after the event has already taken place.”
Thank you for reading! No readers, no point in writing! So if you like my newsletter, I would love, love, love it if you let me know by subscribing to my Substack.
Visit my Website for more: Next Decade
#FutureOfWork #ArtificialIntelligence #Leadership #WorkplaceTrends #FutureReady
This article was originally published in my newsletter, The Next Decade. Read it on Substack or subscribe to get the next edition in your inbox.