The rise of artificial intelligence has been a hot topic for years, with many wondering what the future of work will look like. Will robots take over all our jobs? While that might be a bit of an exaggeration, a recent report from Microsoft Research, detailed in a paper by researchers at Cornell University, gives us a clearer picture of which professions are most likely to be impacted by generative AI.
The paper, titled “Working with AI: Measuring the Occupational Implications of Generative AI,” was authored by Kiran Tomlinson, Sonia Jaffe, Will Wang, Scott Counts, and Siddharth Suri, all from Microsoft Research. The study takes a novel approach to measuring the impact of AI on jobs. Instead of just looking at broad categories, the researchers analyzed the specific tasks involved in thousands of occupations.
They then had AI experts rate how well large language models (LLMs) like GPT-4 could perform those tasks. This allowed them to create a “coverage” score for each job, representing the percentage of tasks that could be automated by AI. This detailed, task-based approach provides a more nuanced understanding of how AI will reshape the workforce.
Key Findings: The Jobs Most and Least at Risk
The report’s findings are both illuminating and, for some, a little concerning. Here’s a breakdown of the jobs most and least likely to be affected by AI, according to the study:
Most Likely to be Replaced by AI:
- Blockchain Engineers: With a 97% coverage score, this is the job the report identifies as most susceptible to AI automation.
- Web and Digital Interface Designers: These professionals also have a high coverage score of 97%.
- Technical Writers: With a score of 96%, AI is poised to take over a significant portion of their tasks.
- Mathematicians: This may come as a surprise, but mathematicians also have a high coverage score of 96%.
- Other high-risk jobs include: Social media managers (95%), proofreaders and copy markers (95%), and accountants and auditors (94%).
Least Likely to be Replaced by AI:
On the other end of the spectrum, jobs that require more hands-on, physical work are less likely to be impacted by AI in the near future. These include:
- Pourers and Casters, Metal: With a 0% coverage score, this job is the least likely to be affected.
- Butchers and Meat Cutters: This profession has a very low coverage score of just 1%.
- Stone Cutters and Carvers: Another hands-on job with a low score of 1%.
- Other low-risk jobs include: Dredge operators (1%), electrical power-line installers and repairers (1%), and athletes and sports competitors (1%).
The “Why”: It’s All About the Tasks
The reason for this disparity lies in the tasks involved. The study found that jobs with a high proportion of tasks related to language, writing, and coding are most likely to be automated. As the authors state, “we find that occupations with higher wages, higher education requirements, and more reliance on analytical and interpersonal skills have higher AI coverage.” This is because LLMs are particularly adept at these kinds of tasks.
Conversely, jobs that involve physical labor, specialized equipment, and real-world interaction are much harder to automate. The report notes that “occupations in production, construction and extraction, and farming, fishing, and forestry have the lowest coverage.”
The Future of Work in the Age of AI
So, what does this all mean for the future of work? While some jobs may be fully automated, the report suggests that in many cases, AI will act as a tool to augment human capabilities rather than replace them entirely. The authors write, “our results suggest that generative AI will have a substantial impact on the labor market, although the exact nature of this impact is still uncertain.”
The key takeaway is that the world of work is changing, and it’s happening faster than we might have expected. For those in professions with high AI coverage, it’s a good time to start thinking about how to adapt and learn new skills. For those in less-affected fields, it remains crucial to stay informed about the evolving technological landscape. The AI revolution is here, and it’s up to us to decide how we want to shape it.
