Overwhelming majority of jobs could be impacted by artificial intelligence – OpenAI

A research paper published by OpenAI, the company behind ChatGPT, found that around 80 per cent of the US workforce could have at least 10 per cent of their work tasks affected by their chatbot or other similar tools.

Researchers investigated the potential implications of language models on occupations within the US job market shortly after the release of the company’s latest machine learning model, GPT-4.

The study involved asking both humans and language models which occupations were fully exposed to LLM tools.

Excited to share our paper on the theoretical economic impacts of GPTs and GPT-powered software!

Excited that people are reading it, but I've seen a few tweets frame this as a prediction, and I wanted to share a few ways that is not the case.

Paper section here, tweets below. pic.twitter.com/HTv5uPmqGe

— pamela mishkin (@ManlikeMishap) March 20, 2023

Being fully exposed does not necessarily mean that tasks of those occupations can be fully automated, but that it could save time in completing a large share of their tasks.

People approached in the study listed a total of 15 occupations which were fully exposed, which included:

• Mathematicians
• Tax preparers
• Financial quantitative analysts
• Writers and authors
• Web and digital interface designers

Meanwhile, language models listed 86 occupations, which included:

• Mathematicians
• Accountants and auditors
• News analysts, reporters, and journalists
• Legal secretaries and administrative assistants
• Clinical data managers
• Climate change policy analysts

However, researchers did admit that the study held some limitations, since the study assumed that language models can’t make decisions or emulate human judgement/empathy.

The study also failed to account to automating one task might affect the time spent on another: “For instance, automating background research could make decision-making more time-consuming,” commented one of the authors.

“We have observed that most occupations exhibit some degree of exposure to GPTs, with higher-wage occupations generally presenting more tasks with high exposure," the authors of the paper concluded.