Generative AI’s potential labor impacts across workers of different educational attainment

Generative AI (GAI), such as Copilot and ChatGPT, is a transformative technology that can change how many people accomplish tasks and spend time during their workday. In prior research, our team categorized occupations into one of three classifications: those with a propensity to be complemented by GAI (augmented), those with a propensity to have its skills replicated by GAI (disrupted), and those unlikely to be impacted by GAI (insulated). In a new white paper, Mar Carpanelli, Cristian Jara-Figueroa, and I examined how workers with different levels of education may be impacted differently by GAI. The educational levels explored in this paper are high school degree, associate’s degree, bachelor’s degree, or graduate degree.

The labor market was already trending towards augmented occupations and away from disrupted occupations

Evaluating the occupations individuals are in for the half decade preceding introduction of GAI, we found that workers in each educational group have been trending towards augmented and away from disrupted occupations. While it’s still too early to say what will ultimately happen, GAI may only accelerate these dynamics, something we will watch with a keen eye.

The potential impact of GAI on workers depends on the skills they have and use

Education levels are related to the three GAI-impact categories, but the relationship is not straightforward. It is not enough to say that having a degree is more—or less—important with the emergence of GAI. The main factor that affects how workers might be influenced by GAI is the skills they possess and the jobs they perform. We see this through the following trends in our data:

1. Each education group has a sizable share of workers in each of the three GAI-impact classifications.
2. Bachelor’s degree holders are among the most likely to be in both augmented and in disrupted occupations, and least likely to be in insulated. 
3. High school graduates also have high shares of workers in disrupted occupations. 
4. Graduate degree holders hold the highest share in insulated occupations.

The impact of GAI will be felt across all education groups, even if by slightly different levels. For example, US LinkedIn members in jobs in the top quartile of GAI-replicable skills are 2X more likely to hold a bachelor’s degree or higher than those in the bottom quartile of GAI-replicable skills.  The skills used in the job and their exposure to GAI is the real determining factor. Workers of all education levels should seek to develop human skills unlikely to be replicated by GAI. There are jobs available at all educational levels for which this is true.

Predicting future trends from GAI

With an eye toward the  future, simulations of potential forward-looking outcomes are helpful for framing around what might happen. For example, if the replicability of skills has a larger impact than the augmenting of skills in comparison to pre-GAI times, we predict that workers will change jobs more often and move somewhat away from augmented and disrupted occupations and into insulated occupations. There may also be a bit more education separation. For example bachelor’s holders may be even more likely than high school graduates to be in augmented occupations, and high school graduates even more likely than bachelor’s to be in insulated occupations.

On the other hand, if the augmenting of skills is stronger than the replicability of skills, we expect to see workers moving somewhat out of disrupted occupations and into augmented and insulated. Workers will be less likely to transition into not working, as demand for workers increases with higher productivity. While each education group benefits from this situation, higher education workers would benefit slightly more than lower education workers, in terms of shifting more towards augmented occupations, as well as decreases in the rate at which they stop working.

Bottom-line implications: GAI will have impacts that cut across education groups, and it’s the skills that matter

Higher education alone does not guarantee disruption or insulation from the impacts of GAI, and not having a degree does not preclude workers from participating in the benefits of GAI in their work. Indeed, the higher potential exposure to GAI that bachelor’s degree holders have over associate degree holders for example (33.4% in disrupted occupations for bachelor’s degree holders compared to 31.7% of associate degree holders) does not imply that all new high school graduates should abandon plans for bachelor’s degrees. Even as we continue to move towards skills-first hiring, college still provides many occupations the best pathway to developing certain in-demand skills, and many college graduates enter into occupations which are likely to be augmented by GAI (indeed, the highest share of any education group). Workers—including new potential entrants to the labor market—should view GAI for what it can be—a complementary technology which can increase productivity of workers in many fields. People skills are one example of this, as are other creative and GAI-supported skills. By understanding the skills best complemented by GAI and pursuing training that opens doors into occupations likely to see increases in demand, workers can better prepare themselves for what the workforce will look like next year and into the future.