Effect of AI on the US labour market

This March 2023 paper investigates the implications of large language models on the U.S. labour market, focusing on how large language model-powered software enhances capabilities compared to large language models alone.

The paper suggests that the impact of LLMs will be substantial, potentially affecting a wide range of occupations and industries.

GPTs are GPTs: An Early Look at the Labor Market Impact Potential...arXiv.org

An Early Look at the Labour Market Impact Potential of Large Language Models

Key findings include

• About 80% of the U.S. workforce could have at least 10% of their work tasks affected by LLMs, with roughly 19% potentially seeing at least 50% of their tasks impacted.

• The effects span all wage levels, suggesting higher-income jobs might face greater exposure to LLM capabilities and LLM-powered software.

• The study indicates that LLMs, particularly when enhanced with software and tooling, could complete 15% to 56% of all worker tasks more efficiently without sacrificing quality.

Conclusion

This study provides a comprehensive analysis of how Large Language Models (LLMs) might influence various jobs and sectors within the U.S. economy.

It introduces a new framework to assess LLM capabilities and their potential impact on employment, revealing that a significant portion of occupations, especially higher-wage ones, are susceptible to LLM integration.

About 19% of jobs could see over half of their tasks affected by LLMs, considering both present and forthcoming LLM-enhanced software.

The findings underscore the widespread influence LLMs could have across numerous professions in the U.S., propelled by further advancements in LLM-driven applications.

While the technical potential for LLMs to augment or streamline human labour is clear, the actual impact on labour productivity will be shaped by a variety of factors including societal, economic, and regulatory considerations.

As LLM technology progresses, its enduring and expanding effect on the economic landscape will present ongoing challenges for policymakers aiming to foresee and manage its implications.

Future research should look into the broader consequences of LLM advancements on the workforce, such as the potential for job augmentation or displacement, effects on job quality, implications for inequality and skill development, and other pertinent issues.

Understanding the capabilities and likely impacts of LLMs will enable policymakers and stakeholders to make well-informed choices to steer the complex interplay of AI with the future of work.