AI Ethics & Fair Housing

Responsible AI principles

• Grounded: the assistant references real public data we have cached for each location.
• Uncertainty-aware: outputs include confidence indicators when interpretation depends on partial data.
• Non-deterministic: we avoid statements like “you will” or “this will happen.” We describe directions and probabilities.
• Non-manipulative: we avoid fear-based or coercive language designed to push a decision.
• Transparent: AI responses are clearly identified as AI-assisted summaries, not human advice.

What our AI will not do

• Tell you that you should or should not live in a specific area.
• Recommend or discourage neighborhoods based on race, color, religion, national origin, sex, familial status, disability, or other protected characteristics.
• Fabricate statistics, forecasts, or sources that are not present in our cached datasets.
• Replace advice from a licensed real-estate, legal, financial, or medical professional.

Fair Housing alignment

We support the principles of the Fair Housing Act. Our scoring, ranking, comparison, and recommendation systems are designed to evaluate places on objective, neutral attributes — affordability, livability, safety, schools, environment, growth — and on the preferences each user voluntarily provides.

We do not allow filters, prompts, or queries that would result in exclusionary recommendations based on protected characteristics. If you believe an output violates these principles, please report it so we can investigate.

Forecasting ethics

Forecasts are presented as directional outlooks with explicit confidence bands. We avoid framing that could create unwarranted urgency, fear, or speculation. Investment signals are interpretive and accompanied by upside, risk, and confidence breakdowns — never as a buy/sell recommendation.

Human oversight

Our team reviews assistant prompts, recommendation logic, and forecast templates for accuracy, neutrality, and Fair Housing compliance. We continuously refine our systems based on user feedback and emerging best practices.