AI in housing

The US experience

In March 2025, a bill was proposed in the¹ state of Georgia (USA) that would prohibit the use of AI in the housing practices if (1) the generation of decisions to sell, rent, finance housing (e.g., mortgages), and (2) the provision of intermediary services (e.g., platform services) or related services (e.g., the delivery of AI- enabled applications to sell or rent housing) occurs without human intervention - without a responsible person reviewing the AI decisions. It is proposed to introduce liability for cases where the user is not warned that an AI or other automated decision tool was used in generating the solutions.

Similar bills have been proposed in other states. For example, as early as 2024, Colorado proposed to recognize as an unfair or deceptive trade practice the use of algorithms in setting² rents amount (bill rejected).³

New York proposed (bill rejected) to prohibit a landlord from using automated decision tools, including AI, unless a developer conducts a disparate impact analysis to assess the actual impact of any automated decision tool at least once a year. The results of the impact assessment shall be made publicly available on the website of the landlord prior to the use of such tool.

Impact assessment includes testing the extent to which the use of an automated decision tool may result in adverse impacts on a group of individuals based on gender, race, ethnicity, and so forth. The landlord should notify any applicant about the use of AI, any such applicant characteristics that the tool will take into account when making rental decisions, provide information about the type, source of data used, and the landlord's data retention policy. If an application for housing is denied, the reason for such denial.

New York is currently considering Bill A⁴ 10020 regarding pricing algorithms - any computational process, including AI, that processes data to recommend prices or terms in the housing practices. It is proposed to prohibit algorithmic pricing tools that use or have been trained on competitors' private data from being used in setting rents. Such data includes nonpublic data that may be shared by landlords, for example, if they use the same application to set housing prices, providing information on housing costs, housing quality (number of rooms, condition), previous rental housing contracts.

Much of this regulation and the attempt to limit the use of AI is due to the risks of anticompetitive behavior and price collusion that drive up housing prices. For example, due to the use of AI-enabled apps that help landlords set housing prices, prices in the US rental real estate market increased by 20% from 2020 to 2024.

Russia’s experience

In Russia, AI is used by housing search platforms (e.g. on Cian), but there is no specific regulation.

From the monitoring issue No. 3 (15), March 2025. Download the full issue (PDF) · issue page at the Gaidar Institute