Predicting AI Job Exposure (www.ben-evans.com)

A recent discussion highlights the challenges in predicting job exposure to AI technologies. Efforts to analyze which jobs and industries are at risk from AI have yielded mixed results, showing that previous waves of technology often disrupted jobs in unexpected ways. For instance, despite significant automation in accounting over the decades, the number of accountants has continued to grow, underscoring that automation can sometimes lead to new job creation rather than outright job loss. This phenomenon is influenced by various factors, including regulatory changes and the so-called Jevons paradox, where cheaper and faster technologies can lead to increased demand for the same tasks. The debate emphasizes the inherent unpredictability in forecasting the impact of AI on employment. Previous technological advancements, such as the internet, transformed business models without necessarily altering the core job functions. The article urges caution in making specific predictions about job displacement due to AI, suggesting that many jobs may experience changes that are not captured by conventional analyses. The complexities of job roles and the interconnectedness of various industries complicate efforts to create accurate models, leading to the conclusion that understanding AI’s effects on employment requires a nuanced approach that considers both direct and indirect impacts.