Stanford University


Suitability for Machine Learning Rubric (WorldSML)


Erik Brynjolfsson
Tom Mitchell
Iyad Rahwan
Sarah H. Bana
Daniel Rock
Morgan Frank


New machine learning techniques and automation are transforming the American workforce, and firms are at a critical juncture. In order to thrive in the digital age, they need to transform their practices and their workforce for the digital world. In previous research, we evaluated every job task in the O*NET database for its suitability for machine learning, or SML, using a rubric that scores each task on 23 different criteria (Brynjolfsson and Mitchell 2017; Brynjolfsson, Mitchell and Rock, 2018). Aggregating these measures to the job level and analyzing the heterogeneity both within and between jobs, this project offers a theoretical framework for how occupations will change and predicts which occupations specifically are most exposed to advances in machine learning methods as they propagate through the network of job tasks. We will apply the rubric to 10 primary sectors of the economy. This approach will provide a broad and deep picture of how technologies such as AI, machine learning, and automation are affecting the workforce, and will be a primary source for informing a new multi-stakeholder approach for the modern global enterprise and also contribute to a public policy manifesto. We will also develop similar rubrics for other technologies, such as robotics, and map these rubrics to the O*NET Database and other data sets.

Related Research

Stanford University