Which jobs will AI replace? Study reveals most at-risk and safest occupations

CAMBRIDGE, United Kingdom — The question of how AI will reshape the workforce has long been a topic of speculation. Now, a new study provides new insights into the types of occupations that may be most impacted by AI technologies.

Researchers from Nokia Bell Labs have developed a novel approach to assess AI’s potential impact on various occupations in the United States. Their findings, published in PNAS Nexus, reveal that AI’s reach extends beyond routine tasks, potentially affecting a wide range of professions, including those traditionally considered safe from automation. (Full list at end of article)

The study introduces a new measure called the Artificial Intelligence Impact (AII) score, which gauges how closely an occupation’s tasks align with recent AI innovations. By analyzing over 24,000 AI-related patents filed between 2015 and 2022, the researchers were able to identify which occupations might be most affected by emerging AI technologies.

Surprisingly, some of the occupations with the highest AII scores were white-collar jobs requiring advanced education and specialized skills. Topping the list were cardiovascular technologists and technicians, sound engineering technicians, and nuclear medicine technologists. Other high-scoring occupations included air traffic controllers, magnetic resonance imaging (MRI) technologists, and even neurologists.

This finding challenges the common belief that AI primarily threatens routine, low-skilled jobs. Instead, the study suggests that AI’s impact may be more closely tied to specific skills within occupations, regardless of whether those skills are considered routine or non-routine.

For example, in healthcare, the study found that 60% of tasks performed by cardiovascular technologists and 48% of those done by MRI technologists could potentially be impacted by AI. Patents related to automating health records management and analyzing medical scans were particularly relevant to these professions.

In the information technology sector, 47% of software developers’ tasks and 40% of computer programmers’ tasks were found to align closely with recent AI patents. These patents focused on automating programming tasks and developing workflows, suggesting that even highly skilled tech jobs may not be immune to AI’s influence.

The manufacturing sector also showed significant potential for AI impact. The study found that 45% of tasks performed by industrial truck and tractor operators and 40% of earth drillers’ tasks aligned with recent AI patents. These patents covered areas such as automated planning for water-well drilling rigs and the operation of electric-powered trucks.

However, the researchers emphasize that a high AII score doesn’t necessarily mean these jobs will disappear. In many cases, AI is more likely to augment human capabilities rather than replace workers entirely. For instance, while AI might automate certain aspects of a neurologist’s job, such as analyzing brain scans, it’s unlikely to replicate the complex decision-making and patient interaction skills that are crucial to the profession.

The study also identified occupations least likely to be impacted by AI in the near future. These tended to be blue-collar jobs requiring physical labor or manual dexterity, such as pile driver operators, dredge operators, and aircraft cargo handling supervisors. Jobs in construction, accommodation and food services, and certain managerial positions also showed low AII scores.

Interestingly, the researchers found a positive correlation between an industry’s AII score and its job vacancy rates. This suggests that sectors potentially most impacted by AI, such as healthcare and information technology, are also experiencing labor shortages. This dynamic could mean that AI adoption in these fields may help address workforce gaps rather than lead to widespread job losses.

As AI continues to evolve, its impact on the job market will likely be complex and multifaceted. While some tasks within occupations may be automated, new roles and responsibilities may also emerge. The key for workers and policymakers will be to adapt to these changes, focusing on skills that complement AI rather than compete with it.