The Allocation of Decision Authority to Human and Artificial Intelligence

Athey, Susan; Bryan, Kevin; Gans, Joshua S.

doi:10.1257/pandp.20201034

Cited by 44 publications

(16 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, when the role of AI as a prediction technology is accounted for, its effects on the labor market are more nuanced (Agrawal et al., 2019a) and not limited to their impact in terms of jobs' destruction. On the one hand, a substitution effect may still arise, as AI may directly substitute capital for labor in prediction tasks, and even in some decision tasks (specifically, when automating prediction increases the relative returns to capital versus labor), raising an issue of organizational design related to the optimal allocation of decision authority to the human rather than to the machine (Athey et al., 2020). On the other hand, AI might enhance labor when automating the prediction tasks, thus increasing labor productivity.…”

Section: Ai and Firmsmentioning

confidence: 99%

Artificial intelligence, firms and consumer behavior: A survey

Abrardi

Cambini

Rondi

2021

Journal of Economic Surveys

View full text Add to dashboard Cite

The current advances in Artificial Intelligence (AI) are likely to have profound economic implications and bring about new trade-offs, thereby posing new challenges from a policymaking point of view. What is the impact of these technologies on the labor market and firms? Will algorithms reduce consumers' biases or will they rather originate new ones? How competition will be affected by AI-powered agents? This study is a first attempt to survey the growing literature on the multi-faceted economic effects of the recent technological advances in AI that involve machine learning applications. We first review research on the implications of AI on firms, focusing on its impact on labor market, productivity, skill composition and innovation. Then we examine how AI contributes to shaping consumer behavior and market competition. We conclude by discussing how public policies can deal with the radical changes that AI is already producing and is going to generate in the future for firms and consumers.

show abstract

Section: Ai and Firmsmentioning

confidence: 99%

Artificial intelligence, firms and consumer behavior: A survey

Abrardi

Cambini

Rondi

2021

Journal of Economic Surveys

View full text Add to dashboard Cite

show abstract

“…2 Understanding AI, therefore, requires not simply asking the aggregate questions of substitution between humans and machines but also of the detailed interaction between humans and AI tools where different aspects of the production function can be either augmented or replaced by technology. Whether and how technology complements or substitutes for skilled expertise will shape both product markets where expert recommenders play a role and the associated labor markets for that expertise (Acemoglu and Restrepo (2019); Athey et al (2020)).…”

Section: Introductionmentioning

confidence: 99%

Managing Intelligence: Skilled Experts and AI in Markets for Complex Products

Gruber¹,

Handel²,

Kina³

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Our work contributes to a nascent literature that studies algorithmic decision-making (e.g. Athey et al, 2020b) and how to regulate it. Most of the existing work in this area has focused on questions of algorithmic fairness, such as work by Gillis and Spiess (2019) and Gillis (2020) on the limits and design of algorithmic audits.…”

Section: Introductionmentioning

confidence: 99%

Unpacking the Black Box: Regulating Algorithmic Decisions

Blattner¹,

Nelson²,

Spiess³

2021

Preprint

View full text Add to dashboard Cite

We characterize optimal oversight of algorithms in a world where an agent designs a complex prediction function but a principal is limited in the amount of information she can learn about the prediction function. We show that limiting agents to prediction functions that are simple enough to be fully transparent is inefficient as long as the bias induced by misalignment between principal's and agent's preferences is small relative to the uncertainty about the true state of the world. Algorithmic audits can improve welfare, but the gains depend on the design of the audit tools. Tools that focus on minimizing overall information loss, the focus of many post-hoc explainer tools, will generally be inefficient since they focus on explaining the average behavior of the prediction function rather than sources of mis-prediction, which matter for welfare-relevant outcomes. Targeted tools that focus on the source of incentive misalignment, e.g., excess false positives or racial disparities, can provide first-best solutions. We provide empirical support for our theoretical findings using an application in consumer lending.

show abstract

The Allocation of Decision Authority to Human and Artificial Intelligence

Cited by 44 publications

References 6 publications

Artificial intelligence, firms and consumer behavior: A survey

Artificial intelligence, firms and consumer behavior: A survey

Managing Intelligence: Skilled Experts and AI in Markets for Complex Products

Unpacking the Black Box: Regulating Algorithmic Decisions

Contact Info

Product

Resources

About