Using Large Language Models to Estimate Valuations of Free Digital Goods, Study Finds Similarities with Human Estimates

Digital goods generate a significant amount of consumer welfare, but many of these welfare gains are not properly measured in official statistics because the goods often lack market prices or are offered to consumers for free. In a new study, researchers investigated the feasibility of using large language models (LLMs) to estimate the valuations of free digital goods by considering a case study estimating valuations of Facebook. They found that valuations generated by LLMs were similar to those estimated using humans and followed similar patterns over time.

The study, by researchers at Carnegie Mellon University, Copenhagen Business School, and Stanford University, appears as a working paper.

“To fully understand the impact of digital goods and track their welfare gains over time, we have to assess how their value has evolved since they were introduced,” explains Avinash Collis, professor of digital economy at Carnegie Mellon’s Heinz College, who coauthored the study. “This is particularly challenging because newly introduced digital goods have limited consumer awareness and current perceptions often bias retrospective estimates of previous consumer valuations.”

LLMs are artificial intelligence systems trained using huge amounts of text data to understand, generate, and manipulate human language. In this study, researchers benchmarked LLMs against valuations obtained from choice experiments on representative samples of U.S. populations from 2016 to 2024. Participants—12,000 in total—were given a single choice between two options—either keeping access to a specific good or giving up access to the good for a period of time in exchange for a monetary offer.

LLM-generated valuations did not differ statistically from valuations obtained through the choice experiments, the study found. In addition, valuations over time followed similar patterns. Moreover, LLMs could be used to extrapolate, going back or forward in time.

“Our approach offers a promising tool for understanding how digital goods like Facebook have evolved in value since their introduction,” says Felix Eggers, professor of marketing at Copenhagen Business School, who coauthored the study. “It also paves the way for extrapolating future valuations and constructing time series data for goods that otherwise lack consistent pricing, including for other non-market goods such as environmental goods.”

Adds Erik Brynjolfsson, professor of economics at Stanford University, who coauthored the study: “The flexibility of LLMs also allows for exploring counterfactual scenarios, such as estimating valuations before a product’s release or future valuations given potential economic or political scenarios, which could provide valuable insights for policymakers and researchers studying the dynamics of digital markets.

“Overall, our findings suggest that LLMs can serve as an effective ‘time machine’ for valuing digital goods, offering both a methodological innovation and a practical resource for longitudinal economic analysis.”

                                                                       ###

Summarized from a working paper, LLM Time Machines: Valuing Digital Goods Over Time, by Collis, A (Carnegie Mellon University), Eggers, F (Copenhagen Business School), and Brynjolfsson, E (Stanford University). Copyright 2025. All rights reserved.

About Carnegie Mellon University's Heinz College of Information Systems and Public Policy
The Heinz College of Information Systems and Public Policy is home to two internationally recognized schools: the School of Information Systems and Management and the School of Public Policy and Management. Heinz College leads at the intersection of people, policy, and technology, with expertise in analytics, artificial intelligence, arts & entertainment, cybersecurity, health care, and public policy. The college offers top-ranked undergraduate, graduate, and executive education certificates in these areas. Our programs are ranked #1 in Information Systems, #1 in Information and Technology Management, #8 in Public Policy Analysis, and #1 in Cybersecurity by U.S. News & World Report. For more information, visit www.heinz.cmu.edu.