Episode 39: Gustavo Augusto Freitas de Lima on Computational Antitrust in Brazil

Thirty antitrust agencies report on their computational work, in what remains the only account written by the agencies themselves. This edition asked them one question. What have these tools allowed you to achieve that you could not have achieved otherwise?Three developments stand out against earlier editions. With no legal regime in place, agencies have begun writing their own rules on what may go into a model. They are putting figures on their work, reporting precision, recall and false-positive rates, which earlier editions almost never did. And what holds deployment back is access to data and the agreements needed to obtain it.Recorded at the Autorité de la concurrence, Paris, 17 September 2026. Open access, 121 pages. shorturl.at/wmEaWSpeakers:➝ Thibault Schrepel, Associate Professor of Law at Vrije Universiteit Amsterdam, Faculty Affiliate at Stanford CodeX, Director of Stanford Computational Antitrust➝ Alba Ribera Martínez, Assistant Professor in Law and Technology at Vrije Universiteit Amsterdam, Editor-in-Chief of Stanford Computational Antitrust➝ Elodie Vandenhende, Deputy Head of the Digital Economy Unit, Autorité de la concurrence➝ Yann Guthmann, Head of the Digital Economy Unit, Autorité de la concurrence➝ Teodora Groza, Postdoctoral Researcher, Chair for Law and Artificial Intelligence, University of Tübingen#antitrust #competitionlaw #artificialintelligence

Can two AI agents fix prices without ever agreeing to? Tanise Brandão and Carlos Neves of Brazi’'s competition agency, CADE, join Thibault Schrepel and Alba Ribera Martínez to walk through their paper on algorithmic collusion.They put two large language models in a repeated Bertrand duopoly, let them price against each other for 400 rounds, and watch supra-competitive prices emerge on their own. The agents coordinate even when they cannot communicate, and some learn to cover their tracks, with one proposing that they steer clear of red flags.The conversation turns to why the choice of model and language moves the result, and why the behavior looks less like textbook rationality than a complex dynamical system. It closes on enforcement, where collusion no longer needs a meeting of minds, and Brandão and Neves make the case for shifting liability from whether firms agreed to what they chose to expose themselves to.https://law.stanford.edu/publications/smart-agent-based-modelling-with-llms-leveraging-large-language-models-for-a-better-understanding-of-algorithmic-collusion/https://www.anthropic.com/research/global-workspace

In episode 39, Alba Ribera Martinez and Thibault Schrepel talk to Gustavo Augusto Freitas de Lima, Interim President of Brazil's Competition Agency (CADE).Follow the Stanford Computational Antitrust project at https://law.stanford.edu/computationalantitrust

In episode 38, Thibault Schrepel talks to Alvin Koh, Chief Executive of the Competition and Consumer Commission of Singapore (CCS).They discuss the creation of the Data and Digital division and the operational role of instruments such as the Complaint Analytics Tool and the AI Verify Toolkit in investigative prioritization and compliance assessment. They also explore the application of computational methods in major enforcement initiatives, including the Grab–Uber merger and the Price Transparency Guidelines, before considering the implications of these practices for the future of competition enforcement in Singapore and the ASEAN region.Follow the Stanford Computational Antitrust project at ⁠https://law.stanford.edu/computationalantitrust

In episode 37, Teodora Groza and Thibault Schrepel talk to Nuno Cunha Rodrigues, President of the Portuguese Competition Authority since 2023. They discuss the tools developed by the agency, how its lawyers work with economists and computer scientists, and the future of the field. Follow the Stanford Computational Antitrust project at https://law.stanford.edu/computationalantitrust

In episode 36, Thibault Schrepel talks to Nancy Scola, longtime technology and politics journalist & reporter.Thibault and Nancy discuss stories that deserve more attention from computational antitrust researchers, how institutions adapt (or fail to adapt) to technological change, and how journalism and computational tools can complement one another. They also consider whether “computational antitrust journalist” could become a job description in the future, and other related topics.Follow the Stanford Computational Antitrust project by subscribing to our newsletter at ⁠https://law.stanford.edu/computationalantitrust

In episode 35, Thibault Schrepel talks to Kamil Nejezchleb, Vice-chairman of Czech Competition Autority.Thibault and Andy Kamil explore how the Czech Competition Authority is integrating computational antitrust into its daily operations. They discuss major cases in which such tools have played a role, the internal structure and expertise needed to support the shift to computational antitrust, the legal constraints imposed by Czech administrative and judicial review, and how the agency envisions the future of enforcement, from algorithmic remedies to cross-border data collaboration.Follow the Stanford Computational Antitrust project by subscribing to our newsletter at ⁠https://law.stanford.edu/computationalantitrust

In episode 34, Thibault Schrepel talks to Andy Chen, Vice Chair (and Acting Chair) of the Taiwan Fair Trade Commission.Thibault and Andy talk about how the TFTC uses computational tools in merger review, price monitoring, and cartel detection. Andy shares insights on the agency’s internal structure, the challenges of explainability, and how computational enforcement might reshape antitrust in Taiwan by 2040.Follow the Stanford Computational Antitrust project by subscribing to our newsletter at https://law.stanford.edu/computationalantitrust

In episode 33, Thibault Schrepel talks to Natalie Harsdorf, Director General of the Austrian Federal Competition Authority. Thibault and Natalie discuss her priorities at the Austrian Competition Authority, with a focus on the role computational antitrust plays in the agency's strategy and daily operations. They also cover how international cooperation in antitrust is evolving and explore what the future might hold for competition policy. Follow the Stanford Computational Antitrust project at https://law.stanford.edu/computationalantitrust.

In episode 32, Thibault Schrepel and Teodora Groza speak with Cary Coglianese, Edward B. Shils Professor of Law and Professor of Political Science at the University of Pennsylvania Law School. They talk about Cary's research on the use of AI in public enforcement. Subscribe to our newsletter at https://law.stanford.edu/computationalantitrust for regular updateson the Stanford Computational Antitrust project.References:- Leashes, Not Guardrails: A Management-Based Approach to Artificial Intelligence Risk Regulation https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5137081- Antitrust by Algorithm https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3985553- Regulating Multifunctionality https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5059426- From Negative to Positive Algorithm Rights https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4225887