Dear Grant County Beat Editor:
NEW MEXICO'S AI DATA CENTERS NEED A LOCAL BENEFITS TEST
Sen. Martin Heinrich's September 17 push to make large AI data centers pay for the grid upgrades they require puts a concrete New Mexico cost question on the table. That principle is necessary but too narrow: communities should also measure whether AI investment builds durable local skills and economic capacity rather than counting construction spending alone. For Grant County Beat readers, the issue connects directly to the state's AI expansion and to Silver City's own effort to help small businesses use AI productively. I propose a local-benefits test that tracks jobs, transferable skills, small-business gains, and the ability of workers to exercise independent judgment around AI systems.
I am a behavioral scientist and the author of the peer-reviewed book, The Psychology of AI Adoption at Work: From Resistance to Results, published by Georgetown University Press https://press.georgetown.edu/Book/The-Psychology-of-AI-Adoption-at-Work My commentary has appeared regularly in The New York Times, The Guardian, the Toronto Star, and many others https://disasteravoidanceexperts.com/media
New Mexico's debate over AI data centers is moving from abstract promise to concrete costs. On September 17, Sen. Martin Heinrich pushed legislation that would require large electricity users such as AI data centers to pay for the grid facilities needed to connect them rather than shifting those costs to families and small businesses.
That is a useful principle. New Mexico should apply the same discipline to the economic benefits these projects promise.
A data center can create construction work, contracts, tax revenue and some long-term technical jobs. But those numbers do not tell communities whether the investment is building durable local capability. The better question is what people and businesses can do after the initial buildout that they could not do before it.
That matters outside Albuquerque and Santa Teresa. Grant County Beat recently highlighted a Silver City workshop designed to help local business owners use AI for practical tasks such as email triage, follow-up and weekly planning. That is a different scale from a hyperscale data center, but it points to the same challenge: the value of AI depends on whether people gain useful skills and retain judgment over the technology.
New Mexico should therefore add a local-benefits test to major AI infrastructure discussions.
First, track durable employment rather than announcing a single headline jobs number. Separate temporary construction work from permanent operations roles, and report how many of those jobs go to New Mexico residents.
Second, measure transferable skill development. Training should leave electricians, technicians, operations staff and local vendors with capabilities they can use across employers rather than skills tied to one facility.
Third, track small-business spillovers. Public officials should know how much local contracting reaches New Mexico firms and whether nearby businesses gain customers, expertise or new markets.
Fourth, measure human judgment. AI systems increasingly influence maintenance, scheduling, security and operational decisions. Workers should be able to verify important recommendations independently and know when they have authority to stop or override an automated system.
This is not an argument against AI infrastructure. It is an argument for judging it the same way Sen. Heinrich proposes judging grid costs: by asking who bears the cost, who receives the benefit and whether the commitments can be measured.
New Mexico is positioned to attract substantial AI investment. The state will get more durable value from that investment if every major project has a public scorecard for jobs, skills, local business gains and human oversight.
---
Gleb Tsipursky, PhD, a behavioral scientist, CEO of Disaster Avoidance Experts, and author of The Psychology of AI Adoption at Work: From Resistance to Results (Georgetown University Press, 2026). https://disasteravoidanceexperts.com/aibook
