Target Audience

Real estate professionals, AI developers, and policymakers concerned with fairness in real estate data and AI systems.

Challenge

Real estate data, particularly in listing descriptions, can inadvertently reflect historical inequalities and discrimination, especially related to race. AI systems using this data risk amplifying biases if not designed responsibly, potentially violating Fair Housing regulations.

Solution Approach

Zillow used NLP and LLM technologies to analyze listing descriptions from homes in neighborhoods with distinct racial demographics (majority non-Hispanic white vs. majority Black). They applied text statistics, key phrase patterns, and topic analysis to identify disparities in semantic meanings, helping to uncover potential biases in the data.

Value Add

This approach enables Zillow to proactively identify and mitigate bias in AI-driven real estate tools, ensuring fairer outcomes for all users while adhering to responsible AI practices and regulatory requirements.

How to Apply This

You don't need months-long projects or huge investments to create an own AI that improves your texts - may it be for reducing bias, optimizing the style and tone to your target audience, and other tasks. What you need is:

• •a clear understanding of what you want to achieve,
• •topic experts that can spend some hours to drive a project,
• •good and bad data that can be used to optimize an AI,
• •guidance by an AI expert to run the AI fine-tuning and set up everything to use your models.

Interested to fighting bias in your own data, or generally improving your content with AI, but unsure about what AI can do for you? Reach out to me and book a free 30-minute feasibility call.

References

Zillow, a leading real estate marketplace, implemented this use case to audit their AI systems for fairness.

Read more here: https://www.zillow.com/tech/using-ai-to-understand-the-complexities-and-pitfalls-of-real-estate-data/