The jobs most at risk from AI

New study pinpoints the job roles that could be replicated by AI. Look away now if you’re a teacher.

Warnings about our jobs being replaced by robots have been around for generations, but developments in technology ‒ particularly in artificial intelligence (AI) ‒ have broadened out the types of job that might be at risk.

A new study from researchers at Princeton University looks to work out the specific jobs most threatened by AI, mapping 10 applications which are powered through AI against 52 human abilities, such as oral comprehensive, oral expression and inductive reasoning.

These capabilities are used by the Occupational Information Network database in the US, which is used by the Department of Labor to describe hundreds of different occupations.

The idea is that if the AI applications can perform those abilities, then they may be able to fulfil job roles where those abilities are key.

The AI apps included features like image recognition, visual question answering, image generation, language modelling, translation and speech recognition.

The most at-risk jobs

These are the ten job roles that the researchers suggested are most at risk from AI:

In fact, across the top 20 at-risk job roles pinpointed by the researchers, 14 are teaching roles.

This is an interesting development from the last time a similar study was carried out, which found that finance roles were most likely to be at risk from AI.

However, what has changed since then has been greater use of language modelling AI, such as ChatGPT.

Why some jobs are at risk

First place though goes to telemarketers.

This perhaps won’t come as a huge shock ‒ many of us will have been presented with some form of ‘chatbot’ when visiting a retail website, which aims to get a better sense of what we might be interested in purchasing, whether it’s an individual item or a service.

The chatbot gathers information about us and then either points us in the right direction, or hands us over to a human salesperson who can help, but this all comes down to the improvements in analysing our responses to those initial questions.

While there has been plenty of discussion around AI ‒ and particularly ChatGPT ‒ in education recently, it has tended to focus on the students themselves.

There have been concerns that this technology makes tasks like setting essay and report-writing tasks a waste of time, since students can swiftly compile a serviceable piece of work by utilising some form of AI app.

However, the Princeton researchers suggest that the role of the teacher could also be taken over by AI, with the report noting that educational occupations are “likely to be relatively more impacted by language modelling than other occupations”. 

This echoes suggestions from Anthony Seldon, the educator and historian, who has predicted that AI will replace teachers as we know them by 2027 ‒ less than five years’ time.

While it’s difficult to envision a robot holding forth at the front of the classroom, it seems certain that the technology has a role to play at least in supplementing the work of teachers when it comes to setting and analysing the work of students.

What it means for the future

Over the last few decades, we have already seen technology take on a greater role in plenty of manual jobs.

The make-up of a car factory today is rather different than it would have been a generation ago, for example.

However, the developments in AI mean that technology will be able to play a greater role in what people might consider ‘white collar’ jobs.

Taking on some of the responsibilities of jobs that rely on maths, like accounting, are a logical next step but the likes of ChatGPT mean that more language-based roles are more under threat.

Only time will tell how well this technology can replicate these responsibilities, and whether its future lies in supplementing rather than replacing specific job roles.

But what’s clear is that all of us will have far more interactions with AI in the years ahead, whether in our own jobs or as consumers.