Generative AI and applications

Generative AI reduces productivity inequality

In this paper, Noy and Zhang analyze the impact of generative AI on workers’ productivity in existing tasks. They recruited 444 experienced, college-educated marketers, grant writers, consultants, data analysts, human resource professionals, and managers and assigned them to complete two occupation-specific, incentivized writing tasks. The tasks included writing press releases, short reports, analysis plans, and delicate emails. The authors divided participants into two equal groups: the “treatment group” whose members are allowed to use ChatGPT for the second assignment, and the “control” group whose members had to use the LaTeX editor of Overleaf. The quality of the completed assignments was assessed by (blinded) experienced professionals working in the same occupations. They found:

• “The generative writing tool increases the output quality of low-ability workers while reducing their time spent, and it allows high-ability workers to maintain their quality standards while becoming significantly faster.”
• “At the aggregate level, ChatGPT substantially compresses the productivity distribution, reducing inequality.”
• “The experimental evidence suggests that ChatGPT largely substitutes for worker effort rather than complementing workers’ skills, potentially causing a decrease in demand for workers.”

Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence
Authors: Shakked Noy, Whitney Zhang
From: MIT

Generative AI in economics

Can machine learning change the scientific process in economics by automatically generating anomalies for existing theories? This is the question Mullainathan and Rambachan attempt to answer in this paper. They define an “anomaly” as a carefully constructed example that provides clues as to how or why a theory fails empirically and argue that these anomalies play a central role in developing new economic theories. The authors develop an econometric framework for anomaly generation and develop two algorithmic procedures to generate anomalies (if they exist) when provided a formal theory and data that the theory seeks to explain.

Our algorithmic procedures are general since anomalies play an important role across a wide variety of fields in economics. As an illustration, we apply our procedures to generate anomalies for expected utility theory using simulated lottery choice data by an individual who behaves according to cumulative prospect theory. We produce novel anomalies for the independence axiom based on the probability weighting function that to our knowledge have not been noticed before. While this illustration is specific, it is our view that automatic anomaly generation can accelerate the development of new theories.

From Predictive Algorithms to Automatic Generation of Anomalies
Authors: Sendhil Mullainathan, Ashesh Rambachan
From: University of Chicago, MIT