Summary: Historical employment data from past technological shifts indicate AI’s impact on job markets will likely mirror previous innovations. The projected 10,000x increase in AI capabilities by 2027 suggests significant productivity gains, catalyzing economic growth and diversifying employment opportunities across sectors, including more need for UX skills.
Relax, Chicken Little! The AI sky isn’t falling — it’s just getting a major upgrade. Sure, robots might steal your job, but they’ll create ten cooler ones you haven’t even dreamed of yet. Time to ditch the doom and gloom and ride the AI wave!
I keep hearing the lament that AI will take our jobs and cause mass unemployment. These negative feelings are based on the true facts that current AI increases knowledge worker productivity by about 40% on a task level and the expectation that future AI may increase team-level productivity by as much as 1,000% (a factor of ten, meaning that a product that used to require 100 people to create will be built by a team of 10 people in the future).
We also know that raw AI power is expected to grow by a factor of 10,000 from 2023 to 2027, resulting in AI capabilities at the level of the best human experts by 2027.
It’s easy to jump from this data to the conclusion that if we only need 10% of the old staff to build a product, we can fire the 90% who are no longer needed.
However, this last reasoning step is false, and based on the “fixed work fallacy,” which posits that there is only a fixed amount of work that’s needed to be done in the world. In reality, if the cost of developing products drops to 10% of the old cost, more products will be developed, and those designers and developers who do not need to build Product One will be redeployed to build Product Two through Ten. In fact, since declining development costs will be married to rising living standards in society, it’s quite likely that dropping costs to 1/10 of before will result in the demand for more than 10x new things, now that people can afford more.
UX’s Future Is Fine
Specifically for my audience of UX professionals, I don’t believe the often-heard UX Angst. Quality user experience is a luxury good. People don’t like me to say this, but it’s true: if customers are poor, they can’t be too picky in demanding high-quality products. They’ll buy something cheap that gets the job done.
On the other hand, when customers (whether individuals or companies) get richer, they allocate some of their newfound wealth to improving their quality of life, which again often means buying higher-quality products. And good UX is definitely a quality that makes life (or work) more pleasant.
Thus, my expectation is that as AI grows the world economy, it’ll also grow the demand for UX quality. Coupled with the fact that UX will be cheaper to deliver due to AI, we should see a lot more good UX in the future, ensuring more than sufficient demand for our services.
My main advice for UX people who worry about the impact of AI on their careers is:
Stop worrying; it’ll be fine.
However, due to pancaking of UX work, there will be less demand for multi-level UX management, so aim your career trajectory more in the direction of becoming the ultimate expert than on growing the number of people reporting to you. Future teams will be small.
Future teams will ship products at a dizzying pace with higher-velocity AI-driven methods than the currently-popular UX lifecycle. The sooner you gain experience with this way of working, the better you’ll be positioned to prosper in the changing work environment. An immediate action item is to use AI-optimized processes in your daily efforts, such as diamond prompting. But it’s probably more important for you to quit your job if you’re stuck in a slow-moving environment dominated by decel management that resists change. If at all possible, join a fast-moving AI-fueled product team.
Broader Employment Trends
It’s rather narcissistic of me to start this article by discussing the future of UX employment, since even with my most optimistic prediction of 100 million UX professionals worldwide by 2050, we’ll only be 1% of the world population and 2% of the workforce.
In Greek mythology, Narcissus was a handsome guy who liked nothing more than to admire his own reflection in the water (this was before decent mirrors). I admit to being a bit like that when thinking about the UX field. We are good, we’re admirable, we have great prospects. But even with much future growth, we’ll only be 1% of the world. So I’ll turn to the 99% of the people untainted by UX beauty. (Midjourney)
Going beyond the employment prospects for UX staff, the same general arguments apply to general employment. No, AI will not cause unemployment, but it will require many workers to retrain for the many new jobs that will be created. Luckily, AI is great for training and serves as a seniority accelerator, so this upskilling can be accomplished if we invest in developing AI for training and education. (Sadly, this does not currently seem to be a priority.)
I wrote an entire article (linked in the previous paragraph) arguing against AI-induced unemployment based on the last 10,000 years of human history. I personally think that when something has been the same for 10,000 years, it’s likely to remain true for the next 100 years, which will only add another percent to that timeline. And it has demonstrably been true for 10,000 years that even the greatest upheavals in the economy, such as the agricultural revolution and the industrial revolution, didn’t create unemployment. For 10,000 years, the people who were no longer needed to perform the old jobs (first hunter, later farmer) got new jobs (first farming, later manufacturing).
When the Agricultural Revolution changed the economy from hunting (panel 1) to farming (panel 2), we didn’t see 98% of the population become unemployed hunters. They became farmers. When the Industrial Revolution changed the economy from farming (panel 2) to manufacturing (panel 3), we didn’t see 98% of the population become unemployed farmers. They became factory workers. My bet is that the same will happen with the AI Revolution. (Leonardo)
This Time It’s Still Not Different
However, I recently read a great article by Michael R. Strain about why AI won’t cause unemployment. He roughly agrees with the points in my article but claims that historical evidence from the old days won’t convince people in the present-day environment. People are easily seduced by the claim that “this time it’s different.” Instead of the long-term trends I like, Strain presents an avalanche of data from the post-World War II era.
Round and around we go, but human nature doesn’t change. People are often fooled by the claim that “this time it’s different.” Usually, it isn’t. (Ideogram)
First up: a chart of U.S. unemployment rates since World War II: lots of zig-zags matching up with various recessions, but the number has fluctuated between 4% and 10% during this long period of substantial change in technology and the economy, with almost all years recording unemployment rates in the narrow band between 4% and 8%. (The one exception: a 15% unemployment rate recorded during Covid.) This is despite the percentage of the workforce employed in manufacturing dropping from 38% to 8%. Those 30% of workers who were fired from manufacturing are clearly not unemployed, since the unemployment rate never surpassed 10% (except for Covid).
Second: Has there been a drop in middle-class income (defined as households earning between $35K and $100K, adjusted for inflation)? Yes, middle-class incomes dropped from 55% in 1967 to 39% in 2022. But this drop was not because workers got poorer. It’s because they got richer! The percentage of households above $100K (again, adjusted for inflation) increased from 13% in 1967 to 37% in 2022. There are now virtually as many rich people as middle-class people in the United States. (And poor folks making less than $35K, who are maybe the category we should care the most about, dropped from 32% in 1967 to 23% in 2022.)
Conclusion: almost a third of the poor people became middle-class people, and almost half of the middle-class people became rich people. This during only 55 years.
The general argument is that people become better off after technological change eliminates old jobs because new jobs are created. Strain gives us an example from banking: In 1990, there were 500,000 bank tellers in the United States. During the next 20 years, the number of automated teller machines (ATM) increased from 100,000 to 400,000. With 4x the automated tellers, what happened to human tellers? Not fired, more were hired. In 2010, 600,000 human bank tellers were employed. What happened is that their job was redefined from counting dollar bills during withdrawals and deposits (something indeed done better by machine) to focusing on more complex customer relationship tasks.
60% of the jobs recorded in 2018 were in categories that didn’t exist in 1940. New jobs are not a rumor, they are most current jobs, if seen from a WW2 perspective. Similarly, in 20 years, when AI has revolutionized the economy, 60% of the population will likely hold jobs that don’t exist today. The added innovation supported by AI’s strong ideation skills will create these jobs. (And no, I can’t say what they’ll be.)
In the 1940s, the United States had about 400,000 telephone company switchboard operators. By now, more than 99% of these jobs have been destroyed, and not because customers stopped placing phone calls. Today, the telecommunications sector employs almost 900,000 people in the United States (but almost no operators).
A telephone company operator connecting calls at a central office switchboard. There are now approximately zero people holding these jobs. But this operator’s grandchildren are not unemployed. They have new jobs that pay much better. (Midjourney)
To conclude, AI will make us all much richer as it expands GDP. This new wealth will create many more new jobs than those that are destroyed by AI.
AI will also provide a decent education for the first time in history for about 350 million children who currently live in extreme poverty, and either don’t go to school at all or attend schools where the teachers don’t report for work on most days. Once these AI-educated kids grow up, they are poised to become dramatically richer than their parents due to their newfound ability to hold decent-paying jobs.
Here’s an overly metaphorical conclusion to this article, suggested by Claude 3.5: In the tapestry of progress, AI threads weave new patterns. As old jobs fade like autumn leaves, spring buds of opportunity bloom. The river of innovation flows ever onward, carrying us not to unemployment’s shore, but to islands of unforeseen potential.