Accurately assessing property values is essential for ensuring that localities have the revenue to support public services like schools, roads, and law enforcement. However, traditional assessment methods face a litany of problems. Valuations can often be inconsistent and municipalities are typically understaffed and resource-constrained.
While there has been a lot of attention on generative AI systems like ChatGPT, predictive AI models trained on property characteristics, sales data, and market trends are increasingly being adopted to address the core challenges of traditional property tax assessment methods.
As home sale data and tax assessments have become easily accessible, researchers have found systematic regressivity in property tax assessment in the past decade. Lower-value properties are typically over-assessed, while higher-value homes are under-assessed. A 2022 report from the Philadelphia Fed, for example, found that “owners of inexpensive houses pay almost 50% higher effective tax rates than owners of expensive houses.” Research on Atlanta’s property taxes, enabled by modeling from Center for Municipal Finance at the University of Chicago, found that 69 percent of the lowest-value properties in Atlanta are over-assessed, while 32 percent of the highest-value homes are under-assessed.
The American Enterprise Institute
By Will Rinehart
September 29, 2025