AI and Jobs: Has the Inflection Point Arrived? Evidence from an Online Labor Platform (2312.04180v2)

Abstract: The emergence of LLMs has renewed the debate on the important issue of "technology displacement". While prior research has investigated the effect of information technology in general on human labor from a macro perspective, this paper complements the literature by examining the impact of LLMs on freelancers from a micro perspective. Specifically, we leverage the release of ChatGPT to investigate how AI influences freelancers across different online labor markets (OLMs). Employing the Difference-in-Differences method, we discovered two distinct scenarios following ChatGPT's release: 1) the displacement effect of LLMs, featuring reduced work volume and earnings, as is exemplified by the translation & localization OLM; 2) the productivity effect of LLMs, featuring increased work volume and earnings, as is exemplified by the web development OLM. To shed light on the underlying mechanisms, we developed a Cournot-type competition model to highlight the existence of an inflection point for each occupation which separates the timeline of AI progress into a honeymoon phase and a substitution phase. Before AI performance crosses the inflection point, human labor benefits each time AI improves, resulting in the honeymoon phase. However, after AI performance crosses the inflection point, additional AI enhancement hurts human labor. Further analyzing the progression from ChatGPT 3.5 to 4.0, we found three effect scenarios (i.e., productivity to productivity, displacement to displacement, and productivity to displacement), consistent with the inflection point conjecture. Heterogeneous analyses reveal that U.S. web developers tend to benefit more from the release of ChatGPT compared to their counterparts in other regions, and somewhat surprisingly, experienced translators seem more likely to exit the market than less experienced translators after the release of ChatGPT.

Introduction

The interaction between AI and employment has been a subject of great interest and debate. How AI affects job markets and individuals' employment opportunities is a question of considerable economic and societal relevance. This paper presents a comprehensive analysis addressing these points, delivering a conceptual framework and empirical evidence from an online labor platform.

Conceptual Framework

The conceptual underpinnings of AI's impact on jobs revolve around four key determinants: task learnability, statistical resources, computational resources, and learning techniques. Task learnability refers to a combination of task statistical complexity and computational complexity, which are intrinsic to any given task. A three-phase relation between AI and jobs is proposed, categorized as decoupled, honeymoon, and substitution phases. This framework becomes insightful when considering that jobs are essentially combinations of various tasks, and AI's ability to assist or replace human efforts in these tasks is heavily influenced by its current level of intelligence.

Inflection Point Theory

Within the conceptual structure, an economic model introduces the idea of an 'inflection point' for each occupation. Before reaching this point, AI's improvements benefit human workers by enhancing productivity. However, once AI crosses this threshold, further AI advancements may negatively impact workers, often resulting in reduced employment opportunities or earnings within the affected occupational categories. These theoretical predictions form the basis for empirical testing.

Empirical Findings

The empirical investigation focused on the effects of a significant AI innovation—the launch of ChatGPT—on two specific job categories on an expansive online labor platform: translation and web development. The findings revealed that translators were negatively impacted post-ChatGPT in terms of job acceptance and earnings, indicating that the occupation of translation might have surpassed the inflection point. In contrast, web developers benefited, with increased job volume and earnings, suggesting that they are still within the honeymoon phase with AI. These results illustrate the nuanced reality of AI's implications across different occupations.

Conclusions and Implications

The paper concludes with reflections on the broader application of the proposed framework and the importance of further research involving more data and occupations. These insights provide a valuable basis for policymakers, educators, and workers in adapting to the evolving landscape of AI in the labor market. Addressing these dynamics is of paramount importance for future workforce development and the sustainable integration of AI into the fabric of the economy.