We often hear that AI will take our jobs. But what jobs will it create?

“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back in 1942, it essentially says that change is inescapable, and that old ways of doing things are constantly being superseded by new ones.

This inevitably means that old jobs disappear, but as creative destruction is powered by innovation, it also means that new jobs are created. It’s no secret that the biggest innovation of this decade is widely expected to be artificial intelligence (AI), so what jobs could AI create.

There are three key areas where job creation could happen according to a new World Economic Forum white paper Jobs of Tomorrow: Large Language Models and Jobs. These are “trainers”, “explainers” and “sustainers”:

Trainers are mainly the people developing AI. This includes engineers and scientists working on the large language models (LLMs) on which generative AI tools such as ChatGPT depend.

But specific roles in this area don’t just belong to programmers designing more efficient algorithms. Electrical engineers could see a rising number of opportunities thanks to demand for customized microchips to train and run LLMs, the Jobs of Tomorrow report says.

Meta is among the companies investigating custom chips for AI. And these more advanced chips could in turn start helping chip designers make further advances, according to Deloitte. It says chipmakers have created AI design tools that are already being used to make chips potentially worth billions of dollars a year.

“Though they won’t replace human designers, their complementary strengths in speed and cost-effectiveness give chipmakers much stronger design capabilities,” Deloitte adds.

Other “trainer” jobs AI could create include systems administrators who are building server infrastructure.

Ensuring these systems have the energy they need will also be crucial, so roles in power systems engineering could also be on the rise, the Forum says.

While trainers are doing the behind-the-scenes work on AI, explainers will be the people making AI easy to use for members of the public.

Explainers will design the interfaces that enable people to interact with AI. They can be thought of as “user experience designers” for LLMs, the Forum says.

People old enough to remember the first PCs may recall having to type in a series of precise technical commands into MS-DOS to get their computer to start up. In a similar way, many of today’s LMMs are still highly technical, meaning they will need well-designed interfaces to make them accessible by the general public, the Forum says.

Explainers could be involved in making LLMs work with different kinds of user inputs. Some may work with typed commands, while others will respond to the spoken voice.

Other explainer work could involve creating LLMs tailored to particular tasks. This could lead to the development of personalized AI assistants, tutors or coaches, the Jobs of Tomorrow report says.

Sustainers will essentially make sure that AI systems are being used in the best way possible. There are likely to be three main types of sustainers, according to the Forum report: content creators, data curators, and ethics and governance specialists.

Content creators – Prompt engineering is a new discipline that involves writing text prompts to make an LLM produce the content a user wants. In the same way the words we type into Google dictate the search results we receive, LLMs need particular prompts to produce the required results.

Optimizing prompts will allow AI sustainers to rapidly produce in-depth content on various topics, in any field or domain, the Forum says. AI has the potential to produce articles, books, teaching and training materials, and even entire storylines for movies and television series – a fact that recently led to strikes by screenwriters in Hollywood.

Data curators – Prompt engineering shows that what you put into AI is key to what you get out of it. Just as important an input for LLMs is the large data sets they are trained on. If the data is not high quality, the output will not be high quality.

Data curators will be tasked with making sure LLMs have the best data going into them. “As most training data are curated from text posted to the internet, data quality and integrity checks are critical, and will lead to the development of its own specialized workforce,” the Jobs of Tomorrow report says.

Ethics and governance specialists – Issues surrounding data quality could also lead to ethical problems in the use of AI. “The presence of prejudiced language in training data can lead LLMs to produce biased, harmful, or unethical content,” the Forum says.

Ethics and governance specialists will be responsible for making sure LLMs do not act in this way. This will involve extensive system testing before anything is released to the public, and could lead to the rise of AI safety officers and ethicists.

A whole new area of regulation and law around AI ethics and governance is also a possibility, leading to new jobs in these areas.

No predictions can be 100% certain regarding which new roles may appear with the widespread adoption of LLMs, the Jobs of Tomorrow report point out. But it sees the highest potential job growth in AI and machine learning specialists in the coming five years, with possible expansion of 39%.

But it also sees growth opportunities in areas such as university and higher education teachers, with 10% growth – a fact also highlighted in the Forum’s Future of Jobs Report 2023.

Jobs such as teaching require a high degree of personal interaction, as do other areas expected to see growth in the years ahead, such as 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 Forum says.

“The adoption of generative AI, particularly LLMs, will transform business and the nature of work, displacing some existing jobs in the process, enhancing others, and ultimately creating many new roles,” the Jobs of Tomorrow report concludes. “It will be incumbent upon business and government to 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 Forum points to the Presidio Recommendations on Responsible Generative AI as a source of potential actions to be taken. It outlines 30 recommendations for the responsible development of generative AI, including practices related to open innovation and collaboration, and priorities for social progress.