Category: Social Disruption

Potential Source of Harm: Labor Market Effects

Updated June 11, 2024

 

Nature of Harm

Over recent years, there have been numerous forecasts that AI-related applications will reduce the number of jobs available to humans. Many of these forecasts have been based upon (apparently) careful research, but they also appear to be driven by (or interpreted to match) the popular / news appetite for bad news about AI.

 

Notwithstanding such predictions, most evidence from previous technological revolutions and so far from AI is that:

  • overall job loss is likely to be low, with the possibility of net job and income gains through economic growth and increasing productivity, but
  • disruption for individuals and groups of workers is likely to be significant due to job changes resulting from progress of AI and related technologies.

 

The International Monetary Fund (IMF) published an excellent summary of research and data on these complex and interacting effects in February 2024: Gen-AI: Artificial Intelligence and the Future of Work. The Massachusetts Institute of Technology has also done excellent work on these issues including Can we Have Pro-Worker AI? (September 2023) and Beyond AI Exposure: Which Tasks are Cost-Effective to Automate with Computer Vision? (January 2024).

 

As one anecdotal example, "buy now pay later" scale-up Klarna reported in March 2024 that it has stopped hiring, including because of workforce efficiencies of AI.

 

There have also been widespread complaints that labelling of training AI data involves low-paid work that exploits workers in developing countries.

 

Regulatory and Governance Solutions

Responding to the major labor market change expected from AI will require a complicated mix of government and private action. The IMF report linked above proposes an AI Preparedness Index on a country-by-country basis, with the following components:

  1. Digital infrastructure
  2. Human capital and labor market policies
  3. Innovation and integration
  4. Regulation and ethics.

 

As the labor market effects of AI increase, government and private sector programs to promote AI preparedness and adjustment will increase. There are also likely to be some direct restrictions on use of AI for certain applications, such as:

  • government and academic institution restrictions on use of ChatGPT and other LLMs – although these have been more because of concerns related to misinformation, academic cheating and infringement than job market effects
  • agreed restrictions on the us of AI in the settlements of the 2023 Hollywood writers and actors strikes
  • restrictions under the EU General Data Protection Regulation on automated processing of personal data (see Infosec - Abuse of Personal Data page).

 

Technical Solutions

There are very few technical solutions to negative labor market effects of AI, because those effects are typically the intended result of the technology rather than an unintended side effect.

 

Government and Private Entities

A large number of government and private entities have been considering labor market effects of AI. A few organizational sites providing resources on this issue are: