OpenAI, the company that developed the popular chatbot ChatGPT, teamed up with researchers at the University of Pennsylvania to examine the long-term impact of AI on the U.S. job market, as well as key job groups. The results are quite disturbing, especially for recent college graduates. The first study of its kind was about finding out which occupations are most likely to be affected by future versions of AI systems such as ChatGPT and may not be needed as much as a result. However, the study not only analyzes the negative sides of the development of AI systems, but also identifies which occupations are likely to be safe from rationalization. This mainly concerns manual labor and skilled trades.
The study found that about 80% of workers in the future should expect that up to 10% of their current work volumes could be performed by a large language learning model (LLM) such as ChatGPT. Furthermore, according to this research, up to 19% of workers can even expect to no longer have to perform up to 50% of their current tasks themselves. The researchers’ assumptions are that if LLMs were widely available, 15% of the total current workload in the U.S. could be done faster by AI tools without sacrificing quality. But the potential of applications such as ChatGPT is far from exhausted: if customized tools and software can be linked to these LLMs in the coming years, as much as between 47% and 56% of all tasks are up for grabs, as AI would complete them faster and more efficiently.
For the complete Study, click here.
