Some researchers are citing the 57% figure as ballpark, based on projections over coming decades. United States numbers in these research scenarios are pretty scary. “Projections” is slightly better than an educated guess, but it’s a big number.
Other researchers cite much lower figures, which doesn’t help clarify anything much. In fact, the plague of predictions about the impact of automation on employment is pretty much all over the place. The reasons for disparity in estimates of this kind are related to how you see, and quantify, impacts on the workforce. Some jobs may go. Some jobs may be created. Sky blue, grass green. Automated everything means fewer jobs, sure, but the slack may be taken up in other areas, human to human jobs like sales, specialist jobs, etc.
Recode.com has a few scenarios which explain the things automation can and can’t do. The actual impact of automation, however, is really rather badly defined, and the history of guesswork is hardly inspiring.
Automation as it now is has been around for a while. Various predicted “end of the world” types of automation are now everyday things. DIY checkouts, ATMs, robot car assembly you name it, there’s a long, decades-old lineage of automation which has replaced many jobs in the last 30+ years. The blue collar jobs went first, the more tech skilled jobs went next. The world, however, did not end.
The global economy “adjusted”, very clumsily, but staggered on in its usual mismanaged delirium. Since then, automation has cranked itself up in to other spheres. There’s even “automated news”, written by software, which is coded to look normal, write to industry standards (terrible thing to say about an innocent bit of code) and deliver basic information.
The changes from these types of basic automation were huge, but their impact was largely overrated. Automation simply reconfigured the workplaces. New jobs did eventually emerge, although the process of adjustment was pretty grim, and brutal to those who lost their jobs. Business downsizing has been notorious for its slash and burn methods, and those methods don’t do much for anyone. Loss of jobs, loss of revenue, loss of income to support business, there was a real shopping list of issues in the transition stage which went largely unnoticed.
Automation + artificial intelligence = What?
The theory of automation vs jobs is based on skills, and that theory is now looking pretty antiquated. The big deal with automation, particularly artificial intelligence governed automation, is what levels of skills automation will have. How many people and roles can it replace? There’s no definitive answer to that, and new tech is already pushing the theoretical horizons a lot farther than most projections.
A major large scale shift to automation would inevitably deliver some shocks, particularly economic shocks. Some areas under scrutiny for the impact of AI, which is itself a form of automation, include:
• Administration
• Accountancy
• Banking
• Corporate management
• Customer relations
• Code writing
• Finance
• Manufacturing
• Mining
• Media
• Retail
• Services sector
Look familiar? These are the sectors where people are being phased out on a generational, not just daily, basis. Workforces shrink, and in these sectors alone, the shrinkage has been huge. It’s not a question of “replacing” people, either; it’s about a totally different business environment.
Automation does deliver efficiency. Mass production was the equivalent of automation in the 20th century. It was very different from the manual work of the previous era, and it was also the basis of the Luddites smashing cotton mills because they took away jobs in the previous century. The problem is that efficiency wins, every time.
Macro economic crash? Not impossible, and potentially very nasty
It’s far too early to say exactly what impact AI + automation will have, let alone produce a blanket figure in a time frame. Quantifying actual scenarios would help, say on a bandwidth, with all jobs lost, or X number of jobs lost, would be more productive.
The problems are likely to emerge as follows:
1. Sudden large scale disruption to incomes over say 5 years
2. Purchasing power is destroyed
3. Debts irrecoverable due to loss of jobs
4. Loss of government revenue through much lower income taxes, etc. leading to higher taxes
5. “Normal life” economics no longer happen; no college, no careers, no mortgage, no jobs anyway.
6. Basic poverty and drastically reduced opportunities for future generations
7. Ridiculous price structures to try to cover reduced revenue for businesses
The inability of a large section of the public to earn a decent living could destroy just about all economic models very quickly, and very thoroughly. The rich would follow down the drain, a bit more slowly, as stagnation causes losses on big financial commitments. This would be like a self-perpetuating Great Depression, with very limited options to support the “human economy”, as it should now be called.
The irony here is that automation and AI will also become obsolete, fast. A lot of money can and will be invested in short term portfolios of tech which have fairly limited shelf lives. In this scenario, automation could theoretically put itself out of a job, but with the economic chaos it caused, nobody could afford to replace it. To quote William Shatner from Airplane 2, irony can be pretty ironic.
