Who Will Pay the Cost for AI, New Data Centers?

By Dr. Gay Lee

Part 2 of a 3 part series

In Part 1, I discussed the environmental and infrastructure costs associated with data centers. But there is another question that concerns me just as much:

Who will actually benefit from the AI economy?

Data centers can create jobs. Building them requires electricians, plumbers, construction workers, equipment operators, HVAC technicians, engineers, fiber-optic workers, and many other skilled professionals.

I welcome those jobs.

But we should be careful about assuming that a billion-dollar data center automatically means hundreds or thousands of permanent jobs.

Construction may require a large workforce. Once construction is finished, the number of permanent employees can be considerably smaller.

In my opinion, companies seeking substantial tax benefits should tell the public exactly what they are promising: how many construction jobs, how many permanent jobs, what those jobs will pay, what benefits will be offered, and when people will be hired.

If public benefits are awarded because a company promises a certain number of jobs and those jobs never materialize, taxpayers should have a way to recover some of those benefits.

Promises should have consequences.

AI Can Create Jobs and Eliminate Them

There is another part of this discussion that I do not believe we can ignore.

Artificial intelligence can increase productivity and create entirely new industries. But companies are also using AI and automation to reduce labor costs and perform tasks previously performed by people.

Some jobs will be created. Some will change. Others may disappear.

My concern is what happens to the worker who has spent 10, 20, or 30 years developing skills in a job that suddenly requires fewer people because of artificial intelligence.

Do we simply tell that person to learn something else?

I don’t believe that is good enough.

The workers most vulnerable to economic disruption may not be technology professionals. They may be administrative workers, customer-service representatives, office employees, production workers, retail employees, and other people whose jobs have traditionally provided a pathway into the middle class.

And lower-income communities may have fewer resources to absorb that disruption.

A community could have a billion-dollar data center sitting nearby while its residents continue struggling with unemployment, underemployment, and poverty.

That is not my definition of economic development.

Economic development should be measured by what happens to people, not simply by the cost of the building.

Training Must Lead to Employment

If companies receive substantial public assistance to build the infrastructure of artificial intelligence, I believe they should also help prepare local workers for the economy they are creating.

Developers should work with unions, community colleges, BOCES, Best Resources, workforce agencies, vocational rehabilitation programs, and community organizations.

We need apprenticeships, pre-apprenticeships, industry credentials, and retraining programs connected to actual employment.
Opportunities should include young adults who do not want a traditional four-year college education, unemployed and underemployed workers, veterans, people with disabilities, and people returning to the workforce.

Workers displaced by technology should also have realistic pathways into growing occupations such as electrical work, cybersecurity, information technology, fiber optics, HVAC, advanced manufacturing, energy systems, and other skilled fields.

My principle is straightforward:

Training should lead to employment.

Public Money Must Produce Public Benefits

I am not opposed to government incentives when they produce real economic development.

But public money should have a public purpose.

If a company receives millions of dollars in tax exemptions, PILOT agreements, grants, or other assistance, taxpayers should know what they are receiving in return.

How many permanent jobs? What wages? What benefits? How much tax revenue? How much workforce training? How many apprenticeships?

And how many local residents will actually have an opportunity to participate?

In my opinion, major technology projects receiving substantial public assistance should enter into a binding Economic Justice, Workforce Development and Community Benefits Agreement.

If we help corporations build the infrastructure of the AI economy, they should help our workers survive and succeed in that economy.
Technology should create pathways into the middle class, not simply another pathway for billion-dollar corporations to become richer.