Rapid advances in artificial intelligence-powered tools are quickly making large language models (LLMs) like OpenAI’s ChatGPT a can’t-live-without resource for many workers.
Executives are relying on the technology to consider different vantage points, real estate pros are using it to write listings and lawyers are using it to generate first drafts of documents like wills and more.
New research examining the effects of language modeling AI like ChatGPT on different occupations and industries finds that certain jobs, like telemarketers and teachers, are more “exposed” to the technology than others, such as psychologists and counselors.
According to the researchers, the term “exposed” has a fairly broad definition, indicating professions may be impacted by AI in a number of ways, ranging from job loss to the inclusion of the technology to assist in some job functions.
Telemarketers topped the list of the 20 most exposed occupations to language modeling AI tools. Postsecondary English language and literature teachers came next, followed by foreign language teachers. Postsecondary teachers of history and other disciplines followed. In fact, teachers claimed nine of the top 11 spots.
These are the 20 most exposed professions:
- English language and literature teachers
- Foreign language and literature teachers
- History teachers
- Law teachers
- Philosophy and religion teachers
- Sociology teachers
- Political science teachers
- Criminal justice and law enforcement teachers
- Social work teachers
- Psychology teachers
- Communications teachers
- Political scientists
- Cultural studies teachers
- Arbitrators, mediators and conciliators
- Judges, magistrate judges and magistrates
- Geography teachers
- Library science teachers
- Clinical, counseling and school psychologists
The legal services industry was found to be the field that’s most exposed to language modeling AI tools. Artist and athlete management, travel agent services and grantmaking and giving services were also among the most exposed fields.
Job augmentation or substitution?
AI researchers from Princeton, the University of Pennsylvania and New York University used the AI Occupational Exposure (AIOE) measure to determine jobs’ and industries’ “exposure” to advances in AI, leaving open the question of whether that exposure leads to augmentation or substitution. They isolated language modeling AI capabilities to hone in on its potential effects on the economy, given ChatGPT’s rapid rise in utilization and popularity.
“We went about it by breaking down occupations by the portfolio of abilities that they require. Different jobs have different skills that are important,” Manav Raj, one of the researchers, told CBS MoneyWatch. “We took an ability-based perspective and measured exposure to advances in different applications of AI at the ability level.”
The researchers took 10 common applications of AI, including image generation like the app DALL-E and language modeling, which powers ChatGPT, and mapped them to different occupations to see how related AI tools’ abilities were to various job skills.
“Given the excitement around ChatGPT and language modeling technologies, we thought it would be interesting to isolate their effects,” Raj said.
Job roles likely to change
As a researcher at a higher educational institution himself, Raj said professors at the University of Pennsylvania are considering how new AI technologies will affect the jobs they do, including how professors go about creating assignments for students.
“There are a lot of really interesting things you can do with ChatGPT, like use it to create a syllabus or suggest readings,” he said.
To him, the overlap makes sense. “The actual abilities educators use on a day-to-day basis involves a lot of content creation and working with language, and that is what these technologies focus on — those abilities,” he explained.
In other words, the nature of work is more likely to change in these occupations exposed to AI.
While in some cases AI will replace workers, in others it will complement their skills and free them up to perform higher-level tasks that boost productivity — and companies’ bottom line. What remains unknown is what kinds of new jobs will be created as a result of emerging forms of AI.
“It’s still unclear exactly what that change will look like,” Raj said. “Maybe it will mean automation and maybe workers will be more productive by using these technologies.”