Up to 80 percent of workers can see jobs being impacted by AI

News

by Isabelle Kane

March 26, 2023 | 9:44 a.m

Artificial intelligence may come to your business.

Researchers from OpenAI and the University of Pennsylvania argued that A New research paper That artificial intelligence may change some areas soon after the advent of ChatGPT, a shockingly intelligent chatbot released in November.

Researchers have argued that 80 percent of the US workforce could be affected by at least 10 percent of their work tasks by introducing ChatGPT.

They also found that about 19 percent of workers may find that at least 50 percent of their duties are affected by GPT, or General Purpose Technologies.

The researchers also found that high-income jobs are likely to have more exposure to GPT, but that it will extend across almost all industries.

The paper examines the “exposure of labor tasks to AI” without distinguishing between the effects of labor augmentation or the effects of labor displacement.

Within the study, the researchers defined “exposure” as a measure of whether access to a GPT system or a GPT powered system would reduce the time it takes a human to perform a work task by at least 50 percent.

The researchers stressed that the exposure does not mean that tasks should be fully automated by GPT, but that the technology can save workers “a significant amount of time to complete a significant part of their tasks.”

The study found that mathematicians, interpreters, accountants, legal secretaries, writers and authors are some of the jobs with the highest levels of exposure.

At the other end of the spectrum, there weren’t a lot of low-paying jobs like railroad maintenance workers, cooks, mechanics, floor layers, meat packers, and masons.

Open AI researcher Pamela Mishkin highlighted the research in a Twitter thread, writing: “Today’s GPTs can do a lot. Over the past few years, we’ve seen them get better and better at solving more and more complex tasks with fewer and fewer examples. on less connected tasks.

She added, “The paper is looking in that direction and not any particular model available today.”