Throughout history, new technology has been a forerunner of better living standards, stronger personal connections and economic prosperity. But these advancements often have unforeseen consequences, especially in the labor market. Today, businesses are exploring ways to integrate artificial intelligence into their operations and understand how that impacts their workforce needs.
According to Pew Research Center, 46% of Americans already use AI tools such as Claude, Copilot and ChatGPT every single day, showing that this technology is already seeing broad adoption.
Just like the internet before it, AI is poised to reshape the way people work, but to what extent is still murky.
READ MORE: How Thomas Couch of Wolverine Builders turned curiosity into a career
Reports from economists differ — Goldman Sachs forecasts AI could automate 25% of all working hours, while Boston Consulting Group projects 12% of current jobs could be replaced entirely.
While estimates of labor market impact vary, one thing is clear: workers across the country want the government to do something about it. A Harris Poll report published earlier this year shows 81% of Americans believe the government should “start preparing programs now” to mitigate job displacement caused by AI.
Figures like these are why researchers at the Arizona Office of Economic Opportunity set out to better understand the effect AI could have on the labor force. The product of their research was the AI Displacement Exposure Index, a model showing which jobs are most at risk of being replaced or downsized due to AI.
Navigating disruption
The methodology used to create this index was wide-ranging, taking into account existing theoretical research from Pew Research Center and the Budget Lab at Yale. It also pulls from labor market data, such as layoffs, new job postings and the capacity of existing workers to adapt to new professions.
According to the OEO’s index, 1.38 million jobs in Arizona are at high or very-high risk of being displaced by AI. These jobs account for 42.9% of the state’s workforce.
“If you use your computer a lot for work, now that same computer can do some of the tasks for you,” explains Eddie Bakko, a senior economist for the OEO.
The jobs at highest-risk are largely white-collar jobs, with over 90% of management, engineering and legal occupations deemed “high-risk.”
Those who don’t need a mouse and keyboard to perform their duties, however, are more likely to be safe from AI automation. These low-risk fields include transportation, personal care and service, and food service.
Even though researchers project high levels of future risk, the OEO found that AI adoption has not yet had a significant effect on current employment figures.
Using late 2022 as a starting point, researchers couldn’t identify specific employment trends directly linked to AI adoption. In fact, employment rates have remained steady for all occupations surveyed, with only a 0.2% decrease in very-high risk jobs measured since 2025.
Bakko notes that the boom in hiring after the COVID-19 pandemic, followed by a decline then leveling out of new job postings from 2023 to 2024, can be attributed to economic factors unrelated to AI adoption.
The future, however, remains less clear — 54% of job postings in 2024 belonged to high- or very-high risk occupations, according to the new index.
Postings in these higher-risk categories also tend to have stricter qualifications. The average experience required for very-high risk fields was 3.4 years, compared to 2.5 years in very-low risk fields. Only 46.8% of job postings in very-high risk fields were entry-level.
This paradigm shift comes with new challenges for Arizona’s labor market. Arielle Amrein, research manager at the OEO, notes that high-exposure fields tend to share skillsets with other high-exposure fields.
To secure their place in the workforce, individuals may need to leave their current professions and retrain for a position in a lower risk field.
Doing so often compresses household budgets, considering that high-exposure occupations earn a median income of $88,249 per year, 76% more than the state’s overall median annual wage.
“We can’t simply advise workers in exposed roles to jump into a lower-exposure job if that means taking a massive pay cut,” Amrein says.
Using this new index, the OEO launched an online portal for members of the public to research how exposed their jobs may be to AI-fueled displacement, and to explore pathways away from high-exposure jobs to other professions.
This portal — aipathways.az.gov — is designed not only to inform Arizona’s workforce, but to point them in the right direction.
On the portal, users can look up their specific job, see recommendations for similar professions with lower exposure and find training/educational programs across the state to retrain into lower exposure roles.
One website cannot solve the effects of AI labor shocks on Arizona’s workforce alone, but it does represent a real first step toward addressing the concerns of workers statewide.
As more companies and industries adopt AI, solutions like these will hopefully soften the blow to those most affected.
