By Zachary Atkins, Staff Writer
Photo courtesy of unsplash.com
With every passing year it seems an especially innovative or extraordinary technological advancement is on the forefront of implementation across our society. Advancements in Artificial Intelligence (“AI”) is no different. AI tools like ChatGPT, Bard, and Dali are reshaping markets with both positive and negative effects on the economy and society. The private sector’s adoption of AI tools necessitates new regulatory approaches and presents opportunities for regulators to leverage digital tools for oversight.[1]
What is artificial intelligence? AI is the development of algorithms and software that enable machines to perform tasks that typically require human intelligence. A commonly utilized branch of AI is machine learning. This is a subset of AI that focuses on the development of algorithms that can learn from and make predictions or decisions based on data.[2] Machine learning allows computers to learn and improve from experience without being explicitly programmed.[3]
Machine learning presents a host of challenges for antitrust enforcement broadly across many different industries.[4] Search engines, e-commerce platforms, and social networks benefit from existing internet economies of scale, and broad access to user driven data. This access and scale are difficult to replicate as a new competitor. These barriers are reinforced by advances in machine learning. It’s important how competition and antitrust enforcement might be enhanced, despite these advantages of incumbent trust activity.[5] Machine learning also challenges antitrust policy by facilitating collusion and price discrimination.[6] Monopolizing platforms pose a large threat to the consumer with their inherent anticompetitive behavior, and present both similar and distinct challenges to regulators as the traditional oil, steel, and sugar trusts posed in the late 19th century.
Antitrust regulators face challenges and opportunities with algorithmic innovations. Machine-learning algorithms can help identify suspicious activities and enhance oversight. Dynamic pricing algorithms introduce complexities and can facilitate anti-competitive behavior.
Prominent platforms such as Google and Facebook earn revenue by matching users with advertisers and products.[7] Google and Facebook display ads alongside organic search results and news feed, respectively, and earn revenue when a user clicks on an ad.[8] Amazon makes money from completed purchases by matching users with product recommendations. The algorithms predict user behavior for clicking on ads or purchasing products, driving the matching process.[9]
Machine learning plays a key role in improving the predictions of the matching algorithms. Alphabet Inc., (Google’s parent company) and Amazon have invested heavily in machine learning, focusing on AI-driven approaches across their products and orienting their operations to optimize accordingly.[10] Amazon uses machine learning for demand forecasting and placement of fulfillment centers. These large firms benefit from the fixed costs of improving algorithms, with access to large data sets giving them an advantage.[11]
Regulators may need to modify existing rules and adopt algorithmic tools for oversight.[12] Other regulators are already using AI and big data to enhance enforcement capabilities.[13]
Transitioning to “antitrust by algorithm” involves strategic decision-making regarding AI tool use and design. Data availability and sharing, real-time data sharing, hardware, and cybersecurity are crucial considerations. Robust internal and external auditing, as well as validating AI tools are necessary for effectiveness and bias mitigation.[14] This must include risk management and independent oversight. There are legal and ethical questions that need to be answered before successful implementation. Embracing “antitrust by algorithm” has the potential to enhance antitrust regulation and enforcement, promoting competition and better consumer outcomes.[15]
Competition against and amongst leading platforms is desirable for society. It encourages lower prices, higher service quality, and greater privacy protection for the consumer. [16] Competition spurs innovation that benefits society, especially from outsiders without existing sales. [17] Incumbent firms may engage in anticompetitive behavior, and policies such as clear labeling of search results, mandated data transparency, and potential legislation to break up leading platforms have been proposed to curb trust activity and increase innovative competition.[18]
[1] Cary Coglianese, AI for the Antitrust Regulator. ProMarket (June, 2023) https://www.promarket.org/2023/06/06/ai-for-the-antitrust-regulator/
[2] A.L. Samuel, Some Studies in Machine Learning Using the Game of Checkers, 3 IBM J. Res. & Dev. 211 (1959) (coining the term “machine learning”)
[3]C. Scott Hemphill. Disruptive Incumbents: Platform Competition in the Age of Machine Learning. Columbia Law Review. (September 2023) https://columbialawreview.org/content/disruptive-incumbents-platform-competition-in-an-age-of-machine-learning/
[4] Maurice E. Stucke, Ariel EzrachiHow Pricing Bots Could Form Cartels and Make Things More Expensive. Harvard Business Review. (October 2016) https://hbr.org/2016/10/how-pricing-bots-could-form-cartels-and-make-things-more-expensive
[5] Ariel Ezrachi & Maurice Stucke, Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy (2016)
[6] Id, Ariel Ezrachi & Maurice Stucke, Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy (2016)
[7] Ajay Agrawal, Joshua Gans & Avi Goldfarb, Prediction Machines: The Simple Economics of Artificial Intelligence (2018)
[8] Id, Ajay Agrawal, Joshua Gans & Avi Goldfarb, Prediction Machines: The Simple Economics of Artificial Intelligence (2018)
[9] Victoria S. Pereira, John E. Susoreny, Derek A. Sutton, Lauren Norris Donahue. Antiturst and AI: US Antitrust Regulators Increasingly Focused on the Potential Anticompetitive Effects of AI. K&L Gates. (September, 2023) https://www.klgates.com/Antitrust-and-AI-US-Antitrust-Regulators-Increasingly-Focused-on-the-Potential-Anticompetitive-Effects-of-AI-9-20-2023
[10] Steven Rosenbush. Big Tech Is Spending Billions on AI Research. Investors Should Keep an Eye Out. Wall Street Journal. (March 2022). https://www.wsj.com/articles/big-tech-is-spending-billions-on-ai-research-investors-should-keep-an-eye-out-11646740800
[11] Robert Zev Mahari, Sandro Claudio Lera, Alex Pentland. Time for a New Antitrust Era: Refocusing Antitrust Law to Invigorate Competition in the 21st Century. Stamford Computational Antitrust. (2021) https://law.stanford.edu/wp-content/uploads/2021/04/pentland-computational-antitrust-project.pdf
[12] Cary Coglianese, AI for the Antitrust Regulator. ProMarket (June, 2023) https://www.promarket.org/2023/06/06/ai-for-the-antitrust-regulator/
[13] Id. (The IRS AI tools to help in detecting tax evasion, while the U.S. Securities and Exchange Commission uses these tools to detect securities fraud.)
[14] Rebecca Crootof. “Cyborg Justice” and the Risk of Technological Legal Lock. Columbia Law Review. (September 2023). https://columbialawreview.org/content/cyborg-justice-and-the-risk-of-technological-legal-lock-in/
[15] Id, Cary Coglianese. AI for the Antitrust Regulator. ProMarket (June, 2023) https://www.promarket.org/2023/06/06/ai-for-the-antitrust-regulator/
[16] Elizabeth Warren, Here’s How We Can Break Up Big Tech, Medium (Mar. 8, 2019), https://medium.com/@teamwarren/heres-how-we-can-break-up-big-tech-9ad9e0da324c [https://perma.cc/4MXA-DN4Z].
[17] Baker. Beyond Schumpeter vs. Arrow: How Antitrust Fosters Innovation, (2007) 74 Antitrust L.J. 575, 577-88
[18] Marvin Ammori, The FTC Should Take a Broader Look at Transparency. Gigaom (June 23, 2012), https://gigaom.com/2012/06/23/the-ftc-should-take-a-broader-look-at-transparency
