As the saying goes, artificial intelligence (AI) is here to stay. Therefore, it is advisable, among other things, to look at which professions are most exposed to it. That is, which occupations currently performed by humans are most likely to be carried out by this technology in the near or medium term. Without a doubt, this information can be interesting for those who are currently considering which profession to pursue. Although it may be surprising, jobs that require a university degree are more exposed.
Read more Carney: “We can create the next world order, one that is fair, stable, and prosperous”
This is pointed out by an analysis from the United States Bureau of Labor Statistics (BLS), reported by Europa Press, which compares 831 occupations according to their employment forecasts for the period 2025-2035 and their relative exposure level to AI.
The research measures to what extent this technology could be used to assist or complete part of the tasks associated with a job. It does not, in any case, imply a forecast of automation, layoffs, or job losses – the study’s authors note – but rather “provides users with additional information to help them make informed decisions about their career planning.”
Of the total (831) professions analyzed, 206 (25%) have a very high exposure to AI; another 206, high; 206 more, moderate; and 213, low. An interesting fact: the results indicate that jobs that usually do not require a university degree are generally less sensitive to AI than those that do. Among the 338 professions analyzed that typically require it, 146 (43%) have very high exposure (aerospace engineers; journalists; accountants and editors; advertising and promotions managers; or anthropologists and archaeologists).
Another 142 (42%) have high exposure (specialized technicians and aerospace engineering and operations technicians; higher education architecture professors; cardiologists; civil engineers; or television, video, and film camera operators).
On the other hand, 39 (12%) are at a moderate level (air traffic controllers; specialized technicians and clinical laboratory technicians; dentists; dermatologists; or forestry and conservation technicians). And only 11 (3%) have low exposure: anesthesiologists; dental hygienists; embalmers; optometrists; or oral and maxillofacial surgeons, among others.
It is worth noting, as mentioned earlier, that higher exposure to AI does not necessarily mean that the profession will have lower employability in the next ten years. In this regard, an occupation requiring a university degree such as aerospace engineer, with a very high exposure level, has an employability increase up to 2035 of 8.3%. Or that of anthropologist and archaeologist, with a growth of 6%.
Something similar happens with occupations with a high exposure level – and also with university education – such as cardiologists (4.8%), civil engineers (6.4%), or higher education architecture professors (2.6%).
The research does not imply a forecast of automation or job losses in any case
However, it is true that there are professions with very high exposure that have a poor outlook. Which ones? Those requiring a university degree, especially professionals in layout and editorial design (-14.5%); basic adult education instructors, secondary adult education, and English as a second language (-13.9%); or computer programmers (-7.3%).
Although, as noted, occupations that do not require higher education are generally less exposed to AI, the top ten professions with the worst employability outlook are dominated by this type of work. The worst affected: word processors and typists (-34.4%). They are followed by telephone operators (-27.6%) and switchboard operators, including call center service (-26%). Of the 493 professions analyzed in this group, only 60 (12%) have a very high exposure level, 64 (13%) high; 167 (34%) moderate; and 202 (41%) low.
Read more The EU proposes limiting social media access to one hour a day for minors aged 13 to 15 years
Jobs requiring a university degree are more exposed to AI
However, it should be clarified that in this top ten there is no unequivocal link between the level of exposure to AI and a poor outlook for professional opportunities. Thus, in five of the ten occupations, the exposure level is low (4) or moderate (1), while in the rest it is high (1) or very high (4).
Just as jobs requiring higher education are, as seen, more exposed to AI, those that do not require it have lower exposure. Of the 493 professions analyzed in this second group, only 60 (12%) have a very high exposure level, 64 (13%) high; 167 (34%) moderate; and 202 (41%) low.
Among the less exposed non-university professions that will increase job creation are massage therapists (15.3%), medical equipment preparers (10.8%), electrical line installers and repairers (10.3%), or aircraft cargo handling supervisors (9.2%). These occupations usually require secondary education or equivalent, or do not require formal academic qualifications.
Likewise, advanced practice nursing professionals, for example, who have university studies, are among the occupations with the highest expected job creation up to 2035, with an increase of 41%, despite having high exposure. This case reflects that AI can transform or complement certain tasks without replacing the demand for professionals.
Also, data scientists (34.6%), medical and health services managers (24.2%), physical therapy assistants (23%), psychiatric technicians (22.3%), computer and information research scientists (21.8%), and healthcare personnel are among those who will generate the most employment.
In parallel, a study by Anthropic published in March indicated that about 30% of workers were excluded from its measurement due to their tasks having too low a frequency in the usage data of Claude to reach the minimum analysis threshold. That group included, among others, cooks, motorcycle mechanics, lifeguards, waiters, dishwashers, and wardrobe attendants.
This company defines exposure as the proportion of tasks in an occupation that AI can accelerate or help complete, not as the proportion of jobs it can replace. And a single job can include potentially automatable functions alongside others that require professional judgment, complex social interaction, or physical activity.
The Anthropic study gives the example of teaching: AI can assist in tasks such as grading exercises, but it cannot manage a classroom on its own. For this reason, some teachers may be less exposed than workers whose functions can be performed entirely remotely.
The company also emphasizes that, so far, higher exposure has not translated into a clear increase in unemployment in those occupations. However, it detects signs of slower hiring among young workers in sectors with high AI exposure.
Read more Mark Carney speaks before the European Parliament and appears alongside Roberta Metsola
