Artificial intelligence could make Michigan businesses dramatically more productive while reducing the number of workers they need. That raises an uncomfortable question: If machines increasingly create the wealth, how do people get enough money to buy what the machines produce?
ANN ARBOR – Yesterday, MITechNews examined three possible artificial intelligence futures for Michigan.
In the most optimistic, AI becomes a powerful assistant that makes workers more productive and creates enough new jobs to replace many it eliminates.
In another, AI becomes a labor substitute. Companies prosper but gradually discover they can accomplish the same work with fewer employees.
Then there is Elon Musk’s extraordinary prediction.
In a recent interview with The Economist, Musk predicted AI could exceed the combined intelligence of humanity within about five years. Combined with advanced robotics, he believes AI could eventually make work optional and produce such extraordinary abundance that money itself becomes largely irrelevant.
Maybe.
But there is an enormous economic gap between today’s Michigan and Musk’s world.
People still need money.
And for most Michigan households, money comes from work.
Who Gets The Productivity Dividend?
Michigan already has roughly 10% of its workforce in occupations considered highly exposed to AI.
Now imagine a Michigan manufacturer discovers AI and automation allow it to produce 20% more with 20% fewer workers.
That’s an extraordinary productivity improvement.
But who benefits?
The company could raise wages. Lower prices. Expand production. Build another factory. Hire workers elsewhere. Increase profits or return more money to shareholders.
Most likely, some combination would occur.
How that productivity dividend gets divided among workers, consumers and owners could become one of the defining economic questions of the AI era.
Because workers aren’t just producing things.
They’re also buying them.
Solution One: Michigan Builds The AI Economy
History provides an optimistic answer.
Technology eliminates jobs but also creates industries.
Automobiles displaced blacksmiths and carriage makers while creating millions of jobs building cars, roads, gasoline stations and dealerships.
AI could do the same.
And Michigan has an advantage many states don’t.
Michigan knows how to manufacture things.
If AI moves from computer screens into the physical world, somebody must manufacture the robots, autonomous machines, sensors, control systems, power electronics and other equipment that intelligent machines require.
Michigan’s automakers, suppliers, engineers, tool-and-die companies and automation specialists could potentially capture part of that enormous industry.
Instead of simply watching AI-powered robots replace Michigan workers, Michigan could design, engineer and manufacture those robots.
That could create jobs for robotics technicians, engineers, electricians, software developers, machinists, cybersecurity specialists, skilled trades and occupations that don’t exist today.
But capturing those jobs raises another challenge.
Michigan workers will need the skills to fill them.
That could mean dramatically expanding technical training, community college programs and traditional earn-while-you-learn union apprenticeship models for an AI economy where demonstrated technology skills may sometimes matter more than a four-year degree.
We’ll examine that challenge Monday.
Solution Two: Work Less Instead Of Eliminating Workers
AI productivity could also become time instead of layoffs.
Suppose AI allows a company to accomplish five days of work in four.
Instead of eliminating 20% of its workforce, it could move toward a four-day workweek.
Workers retain their income while gaining more leisure. Businesses retain experienced employees. Consumer purchasing power remains in the economy.
There is historical precedent.
American workers once routinely worked six-day weeks. Rising productivity helped make the five-day, 40-hour workweek possible.
The obstacle is economics.
A company capable of producing the same output with 800 employees may have little incentive to continue paying 1,000.
Solution Three: Workers Get A Bigger Piece
Companies could share AI productivity gains with remaining employees.
Higher wages, bonuses, employee stock ownership and profit sharing could put some of AI’s additional economic value into workers’ pockets.
Businesses become more productive. Workers earn more. Workers spend more.
But that isn’t guaranteed.
If AI reduces employers’ dependence on human labor, workers could have less bargaining power, allowing more of the gains to flow to the owners of the technology and businesses.
Solution Four: Everything Gets Cheaper
Musk offers another possibility.
People don’t need as much income if the things they buy become dramatically cheaper.
AI and robotics potentially could lower the cost of manufacturing, transportation, software, professional services and many consumer products.
If household income falls 10% while the cost of living falls 30%, a family theoretically becomes better off.
But abundance has limits.
Land remains scarce. Housing requires land and materials. Electricity requires generating plants and transmission lines. Data centers require chips, water and enormous amounts of power.
AI can reduce scarcity.
It cannot necessarily eliminate it.
Solution Five: Who Owns The Machines?
That leads to perhaps the biggest question.
Who owns the AI and robots?
If machines perform an increasing share of Michigan’s economic work, ownership could become as important as employment.
Workers could share in AI-created wealth through employee stock ownership, retirement investments, pension funds, profit sharing or other mechanisms.
If labor becomes less important as a source of income, ownership becomes more important.
But if ownership of AI, robots, data centers and other productive assets becomes concentrated among relatively few companies and individuals, enormous productivity gains could also produce enormous wealth concentration.
Solution Six: Government Steps In
If technology ultimately eliminates enough employment, some form of government redistribution inevitably enters the debate.
Universal basic income is one possibility.
Musk talks instead about “universal high income,” imagining AI and robotics producing such abundance that everyone can enjoy a high standard of living.
But the transition raises the difficult question: Who pays?
Taxes on corporate profits, investment income, wealthy households or AI-generated economic activity are possibilities.
Each would produce major economic and political consequences.
Michigan government could face its own problem.
Its finances depend heavily on people working, earning and spending. Significant declines in employment and labor income could weaken tax revenue just as displaced workers require additional assistance.
Businesses May Decide First
Long before Lansing or Washington redesigns the economy, Michigan businesses will make thousands of individual decisions.
When AI increases productivity, companies can eliminate positions, shorten workweeks, raise wages, lower prices, expand production, invest in new operations or increase profits.
Those decisions may determine how broadly AI prosperity gets distributed.
Historian Yuval Noah Harari’s concern about AI ultimately comes down to a similar question: not simply how intelligent the technology becomes, but who controls it and who benefits.
AI could produce extraordinary prosperity.
It could also concentrate extraordinary wealth.
Michigan has a particular opportunity because it doesn’t just consume technology.
It builds things.
The state could become a major manufacturer of the machines powering the AI economy rather than simply a customer for technology built elsewhere.
But that will require Michigan workers capable of building them.
AI may determine how much wealth Michigan can produce.
It won’t determine who gets it.
Monday: Does Michigan Need A New Apprenticeship System For The AI Age? MITechNews examines whether Michigan’s traditional earn-while-you-learn skilled-trades model could prepare workers for robotics, automation and AI-driven manufacturing.
