180 incidents on record · 2026 Headlights Incident reports by Ellie Harris · Melbourne
10 new this week Library last updated 30 August 2026
← The incident library
HD-INC-154
Housing · United States · 2025 · Algorithmic collusion

The US Department of Justice alleged that competing landlords set rents using a shared pricing algorithm fed with each other's confidential data, and RealPage settled without admitting wrongdoing

By Ellie Harris · Filed Complaint filed 23 August 2024 in the Middle District of North Carolina

Alleged: RealPage, Inc.; landlords named separately including Greystar, Cortland Management, LivCor and Willow Bridge developed or deployed the AI system implicated in this incident. Details are drawn from public reports; parties are presumed innocent of any wrongdoing not established by an official finding.

The US Department of Justice alleged that competing landlords set rents using a shared pricing algorithm fed with each other's confidential data, and RealPage settled without admitting wrongdoing

What happened

It was reported that on 23 August 2024 the United States Department of Justice and eight states filed an antitrust complaint against RealPage in the United States District Court for the Middle District of North Carolina. The states were North Carolina, California, Colorado, Connecticut, Minnesota, Oregon, Tennessee and Washington. RealPage sells software that recommends rents to apartment landlords. The department alleged that users relied on competitors’ competitively sensitive information supplied through the software rather than setting prices independently. These remain allegations and no liability has been established. RealPage told CNBC that its revenue management products make recommendations for roughly 4.5 million United States housing units, and told The Washington Post that about 600 customers use the software for more than 4.5 million residential units. Those company figures describe a product used across a substantial part of the rental market.

It was reported that on 24 November 2025 the department filed a proposed final judgment and announced it the next day. The proposed judgment would prevent RealPage from using competitors’ non-public, competitively sensitive information to determine rents during runtime. It would also restrict model training to backward-looking non-public data at least twelve months old. The department said the settlement was intended to restore competition for millions of renters. RealPage said the agreement contains no financial penalties, damages, findings or admissions of wrongdoing. It remains subject to court approval. Six large landlords were sued separately over the same alleged conduct. Proposed settlements with Cortland Management in January 2025, Greystar in August 2025, LivCor in December 2025 and Willow Bridge in July 2026 imposed similar restrictions on competitively sensitive data and on participation in RealPage-hosted meetings of competitors.

What an auditable version would have shown

The dispute turned on the data used to produce rent recommendations, and establishing that answer required a federal antitrust investigation. An auditable version would retain it within the software. For each recommendation the record would identify every input, its source, whether it was public or non-public, whether it came from the customer or a competitor, and its age. Such a record would not decide whether the conduct was lawful. It would allow a regulator, customer or court to examine the inputs directly instead of reconstructing the product through discovery and expert evidence. It would also show a landlord whether a recommendation relied only on its own figures and public market data, distinguishing analysis of a firm’s own position from a shared system operating between competitors.

Where the gap was

It was reported that the department’s allegation concerned the provenance and age of data used by the pricing model. The proposed remedy addresses the same features by excluding competitors’ non-public information from runtime and requiring at least twelve months of age for non-public training data. A ConductRecord preserves a signed account of each recommendation and its inputs, making provenance part of the file rather than a matter for later reconstruction. A ConstraintGate applies standing rules before each call, refusing and recording an input prohibited because of its source or age. RealPage disputes the department’s characterisation of its software, and no wrongdoing has been found. The proposed controls address the operational facts without determining the unresolved legal allegations.

What governance should have looked like

It was reported that the proposed final judgment imposes rules on data source and age after the software had become part of pricing across a large share of the market. When competitors rely on a shared pricing system, data provenance belongs among its first controls. The software should identify each input’s source, separate public from non-public information, apply any age requirement and preserve those facts with every recommendation. The rules should operate while the system is in use rather than depend on later reconstruction. RealPage denies wrongdoing, the department’s case remains an allegation and the proposed judgment does not establish liability. Those legal questions remain separate from the operational issue the case exposed, that people outside the company could not determine what data had produced a rent recommendation before the proceedings.

Failure Pattern: competing firms took price recommendations from a shared algorithm fed with one another’s confidential data, and treated the output as an ordinary input to pricing.

Governance Principle: where a pricing tool draws on data from competing firms, the source and age of every input should be recorded and constrained by rule, so that what enters a price recommendation can be examined rather than assumed.

The reference implementation of ConductRecord and ConstraintGate is open source. It lives at github.com/saffronandindia/headlights-oss, Apache 2.0 licensed and free to install. The repository is public now.

Sources

The mailing list

Fresh incident reports every week. One email to match.

We add new incidents to the library regularly, and send a single short email each week with what's new. The library stays free and open; this is just how you keep up with it.

No tracking. Unsubscribe in one click.

The record

An auditable system would have produced a signed, tamper-evident record the moment this happened: what the system did, the version that did it, the basis it acted on, and the action taken, and RealPage, Inc.; landlords named separately including Greystar, Cortland Management, LivCor and Willow Bridge could have produced it on demand.

This is the record the system as deployed did not produce in a signed, auditable form.

What this teaches
Capture what happened when it happens
What the system did, the version that did it, the basis it acted on, and the action taken, recorded at the moment, not reconstructed after.
Sign it, so no one has to trust the record-keeper
A tamper-evident entry. Edit it later and the signature breaks. The record does not ask for the benefit of the doubt.
Make it verifiable by anyone
A court, a regulator, a customer's lawyer can check the record themselves, without taking the company, or us, at our word.

Headlights summarises publicly reported AI incidents. All summaries are independently written, attributed to their original sources, and intended for research and educational purposes. Allegations are identified as such until established through official findings.

This report is based on the sources listed above and reflects information available at the time of review; later developments may not be captured. Where a person is described as charged with or alleged to have done something, that allegation is unproven unless a conviction or a court or regulatory finding is stated. Headlights publishes journalism and commentary, not legal advice.

Want to write back?

Direct to my inbox.

ellie@useheadlights.com →