From Occupations to Tasks: A New Perspective on Automatability Prediction Using BERT

TL;DR

This work reframes automation risk assessment by predicting automatability at the task level instead of the occupation level. It trains a BERT-based classifier on task statements from O*NET, ESCO, and AU LMIS with expert-labeled Task Automatability categories (Substitution, Complementarity, Negligibility) and demonstrates superior performance over traditional models and other transformers. Aggregating task-level predictions to occupations reveals that approximately $25.1\%$ of ONET occupations face substantial automation risk, highlighting diverse vulnerability across sectors. The findings offer a robust tool for policymakers, workers, and industry leaders to anticipate automation impacts and guide reskilling decisions amid rapid technological change.

Abstract

As automation technologies continue to advance at an unprecedented rate, concerns about job displacement and the future of work have become increasingly prevalent. While existing research has primarily focused on the potential impact of automation at the occupation level, there has been a lack of investigation into the automatability of individual tasks. This paper addresses this gap by proposing a BERT-based classifier to predict the automatability of tasks in the forthcoming decade at a granular level leveraging the context and semantics information of tasks. We leverage three public datasets: O*NET Task Statements, ESCO Skills, and Australian Labour Market Insights Tasks, and perform expert annotation. Our BERT-based classifier, fine-tuned on our task statement data, demonstrates superior performance over traditional machine learning models, neural network architectures, and other transformer models. Our findings also indicate that approximately 25.1% of occupations within the O*NET database are at substantial risk of automation, with a diverse spectrum of automation vulnerability across sectors. This research provides a robust tool for assessing the future impact of automation on the labor market, offering valuable insights for policymakers, workers, and industry leaders in the face of rapid technological advancement.

Figures (7)

• Figure 1: The framework of the proposed method.
• Figure 2: The Comparison of different augmentation.
• Figure 3: The performance of model at different data split.
• Figure 4: The Wordclouds for different categories.
• Figure 5: The Distribution of O*NET Task Automatability.