Preventing Algorithmic Collusion Act of 2025
- Sponsor
- Sen. Klobuchar, Amy [D-MN]
- Committees
- Judiciary Committee (primary)
- Last action
- Jan 23, 2025
Bottom line
This bill seeks to prevent algorithmic price collusion and increase transparency in automated pricing by restricting data sharing and enhancing regulatory oversight and enforcement capabilities.
What it actually does
This bill prohibits the use or distribution of pricing algorithms that incorporate nonpublic competitor data, establishes a rebuttable presumption of agreement for certain algorithmic price fixing under antitrust laws, mandates transparency disclosures for businesses using pricing algorithms, and creates an audit tool for antitrust agencies to scrutinize these algorithms. It aims to modernize antitrust enforcement to address potential collusion facilitated by artificial intelligence.
Proponents argue
Supporters argue that pricing algorithms, especially those leveraging shared nonpublic competitor data, can facilitate illegal price fixing, harm consumers through artificially inflated prices, and suppress wages. This bill provides necessary tools for regulators to combat these modern forms of collusion, ensure fair competition, and protect consumers and workers in an increasingly algorithm-driven economy.
Opponents contend
Critics may contend that the bill's broad restrictions on data use could stifle innovation in pricing technology, impose significant compliance burdens on businesses, and that existing antitrust laws are already sufficient to address collusive behavior, regardless of the tools used. They might also raise concerns about the legal implications of establishing a presumption of agreement for algorithmic conduct.
The bill is relatively concise and clearly structured, making it accessible for an informed reader to understand its core provisions and implications within a reasonable timeframe.
Section 3
Antitrust Law Enforcement Audit Tool
This section requires any person using or distributing a pricing algorithm to provide a detailed written report to the Attorney General or the Federal Trade Commission (FTC) upon request. The report must include information about the algorithm's development, whether it autonomously sets prices, its operational rules, the data it uses (including training data), its data collection processes, and whether it engages in price discrimination. A senior corporate officer must certify the report's accuracy under penalty of perjury.
Supporters argue
Supporters argue this audit tool is crucial for regulators to understand complex algorithmic pricing practices, identify potential anticompetitive behavior, and effectively enforce antitrust laws in the digital age, where traditional investigative methods may fall short.
Critics contend
Opponents might contend that these reporting requirements are overly broad, burdensome for businesses, and could expose proprietary trade secrets, potentially stifling innovation in algorithm development and competitive strategies.
Tradeoffs
This provision balances the need for robust regulatory oversight in a rapidly evolving technological landscape against concerns about compliance costs for businesses and the protection of intellectual property.
Section 4(a)
Prohibition on Pricing Algorithms Using Nonpublic Competitor Data
This section makes it unlawful for any person to use or distribute a pricing algorithm that incorporates, uses, or was trained with nonpublic data obtained from another competitor in the same or a related market. Violations are subject to civil penalties of not less than $10,000 per day (adjusted for inflation) or the sum of the price of each product or service sold using the offending algorithm, along with other appropriate relief like injunctions.
Section 5(a)
Presumption of Agreement for Algorithmic Price Fixing
This section establishes a rebuttable presumption of an agreement, contract, combination, or conspiracy under the Sherman Act (for restraint of trade) or an unfair method of competition under the FTC Act. This presumption applies if a defendant distributed a prohibited pricing algorithm to two or more persons who then used it to set prices or commercial terms in the same or a related market, or if the defendant used such an algorithm and another person also used it in the same or a related market.
Section 6
Mandatory Disclosures for Algorithmic Pricing
This section requires businesses with $5 million or more in annual revenue that use pricing algorithms to clearly disclose to customers (before purchase) or employees/contractors (before service) that prices or commercial terms are set or recommended by an algorithm. Additional disclosures are mandated if the algorithm engages in price discrimination (setting different prices for similar products/services or wages for similar work) or if it's a third-party algorithm (requiring identification of the developer/distributor). Failure to provide these disclosures constitutes an unfair or deceptive act under the FTC Act, with civil penalties of not less than $5,000 per day.
The definition of 'nonpublic competitor data' is crucial. It means nonpublic data derived from or provided by another person competing in the same or a related market. Importantly, it explicitly *excludes* information distributed, reported, or communicated in a way that does not reveal any underlying data from a competitor, such as narrative industry reports, news reports, business commentaries, or generalized industry survey results.
Section 2(5)
Why it matters:It is standard legislative practice to define terms at the beginning of a bill. The exclusion is a necessary clarification to prevent the main prohibition from being overly broad and to allow for legitimate market research.
Case for: Supporters would argue this precise definition is vital to target genuinely collusive data sharing while allowing legitimate market research and publicly available information to be used, ensuring the bill is appropriately targeted.
Case against: Critics might argue that even with this exclusion, the definition could still be interpreted broadly by regulators, leading to uncertainty for businesses about what data is permissible and potentially chilling legitimate competitive intelligence.
Estimated impact: This definition significantly shapes the scope of the prohibition in Section 4, determining which data uses are legal and which are not, thereby impacting the entire landscape of algorithmic pricing.
This subsection outlines how a defendant can rebut the presumption of agreement established in Section 5(a). The presumption does not apply if the defendant did not develop or distribute the pricing algorithm AND demonstrates by clear and convincing evidence that they did not have actual knowledge or could not have reasonably known that the pricing algorithm used nonpublic competitor data.
Section 5(b)
Why it matters:It is standard legislative practice to include exceptions or rebuttals immediately following the rule they modify, ensuring the legal framework is comprehensive and addresses potential overreach.
Case for: Proponents would argue this provides a necessary safeguard for innocent parties, ensuring that only those truly culpable for facilitating algorithmic collusion are penalized, particularly third-party users of algorithms who may not control the underlying data sources.
Case against: Opponents might argue that 'clear and convincing evidence' is a very high bar, and the provision still places a significant burden on defendants to prove a negative, even if they are not the algorithm's developer or distributor, potentially leading to unfair outcomes.
Estimated impact: This provision crucially impacts the legal defense strategy for businesses accused of algorithmic price fixing, especially those who are users rather than developers or distributors of the algorithms.