AI-proof jobs: a matrix to determine your career

Some workers will be needed to train, explain, and sustain large language or other AI models. But what about the rest of us? For many, the labor market will change beyond recognition, with AI erasing and reshaping various jobs.

23% of global jobs will change in the next five years due to industry transformation, including through artificial intelligence (AI) and processing technologies, the latest World Economic Forum (WEF) white paper on the Jobs of Tomorrow reveals.

ChatGPT and other large language models (LLMs) could erase up to 62% of work time in some tasks and “could significantly impact a broad spectrum of job roles.”

Those in finance, IT, engineering, and media should brace themselves for changes. If they aren’t swept away by automation, they will be “augmented.”

“Creative destruction” is one of the core concepts of economics. Identified by Joseph Schumpeter back in 1942, it essentially says that change is inescapable, and that old ways of doing things are constantly being superseded by new ones,” the WEF writes.

While AI is expected to create new fields of work and lead to large expansion in others, there are no 100% guarantees on AI job creation.

The WEF matrix can help you determine where your profession stands in the expected path of the AI bulldozer in the next five years. Released September 2023, the analysis represents jobs expected to decline or grow, together with the level of “automation” or “augmentation” introduced by LLMs.

The “safest” jobs are those that require a high degree of personal interaction, such as teachers or healthcare professionals.

“Roles that are not reliant on language-related tasks – such as manual work – are also unlikely to be detrimentally impacted by AI,” the WEF says.

How to read this matrix: viewing vertically, the jobs with the highest automation potential are in the lowest boxes, in the middle boxes are jobs with low potential for AI exposure, and the jobs with higher potential for augmentation are in the top boxes. Horizontal distribution indicates net expected growth for jobs within the next five years, measured as the expected percentage change in workforce employment.

Some jobs will change beyond recognition

LLMs have the largest potential to automate jobs that are repetitive and do not require much interpersonal communication.

Credit authorizers, checkers, and clerks have the highest amount of potentially automatable work, 81%, according to WEF. Another 7% of their work is exposed to AI augmentation.

Management analysts are next in line, with 70% of their work exposed to automation. AI seemingly could wipe out more than half of Telemarketers (68% of work exposed to automation), statistical assistants (61%), tellers (60%), forensic science technicians, receptionists, information clerks, accountants, administrators, etc.

The jobs exposed to automation usually overlap significantly with the jobs that are expected to decline over the next five years.

Other professions that will have to learn to work with AI are those exposed to “augmentation.” Those professions require critical thinking and complex problem-solving skills, especially in science, technology, engineering, and mathematics.

Not a single insurance underwriter will have AI-free work, as this occupation is 100% exposed to augmentation, according to WEF analysis. The same goes for bioengineers and biomedical engineers (84% of work exposed to augmentation and 16% to automation).

Mathematicians, editors, database architects or administrators, statisticians, graphic designers, or translators should be aware that similar work is more than 60% exposed to augmentation, and automation will further remove a large chunk of the remaining work time.

“Many jobs with the highest potential for augmentation also have some potential for automation, resulting in very high total exposure for these jobs, such as medical transcriptionists, insurance appraisers, and assessors of Real Estate,” analysts write.

The most AI-affected industries will be financial services and capital markets, insurance, pension management, information and technology services, telecommunications, media and publishing, research design, business management, rental, leasing, and retail.

Some jobs will remain (almost) the same

If you don’t feel like becoming an AI puppet or puppeteer, there’s some good news. There are professions that machines won’t have much effect on in the near term.

These jobs usually require a high degree of personal interaction, such as healthcare professionals or teachers, or physical movement, such as athletes or manual laborers.

So, educational, guidance, career counselors, and advisers will have 84% of their work time left, followed by clergy with a similar AI-proof work amount.

Legal advisors, home health aides, anesthesiologists, social workers, marriage and family therapists, other medical workers, athletes, and teachers – all these will have more than 50% of work time unexposed for AI mingling.

Of course, jobs that require physical work and no language tasks will remain unaffected by AI. These jobs include highway maintenance workers, rail-track layers, various operators, slaughterers and meat packers, dishwashers, fabricators, etc.

The least affected industries will be agriculture, forestry and fishing, production of consumer goods, mining and metals, chemicals, and advanced materials.

No 100% certainty for AI job creation

While AI introduces a new paradigm of collaboration between humans and AI, WEF analysts could not make foolproof predictions regarding what new roles may appear with the widespread adoption of LLMs.

The report sees the highest potential job growth in AI and machine learning specialists in the coming five years. The growth is measured in percentages, with a possible expansion of 39%.

Opportunities will likely rise fastest for AI trainers, explainers, and sustainers.

AI model and prompt engineering is one good guess, as these skill sets evolve simultaneously with AI systems. These jobs cover a range of programmers, electrical engineers, systems administrators, and infrastructure and power systems engineers.

“Prompt Engineers will be critical to developing, refining and reframing prompts or inputs for LLMs to reach more optimal results,” the WEF suggests.

Other emerging jobs include interface and interaction designers, AI content creators, data curators and trainers, and ethics and governance specialists.

“While the emergence of new job categories can be expected, the reinvention of existing roles should also be anticipated,” the analysis reads.

Researchers conclude that LLMs “present an opportunity to extend human potential, grow industries and strengthen global economies. Yet their rapid adoption contains both risks and opportunities for the workforce.”

Both politicians and managers must take proactive steps in preparing the workforce for the “extensive transition to come, to ensure that all members of society benefit from the potential of generative AI.”

The paper provided an analysis of over 19,000 individual tasks across 867 occupations to assess the impact of LLMs on jobs.