The honest answer matters, because housing discrimination is both wrong and against the law. The good news is that it is avoidable.
Why careless AI scoring is dangerous
- An AI trained on a pile of past accept and reject decisions learns to imitate them, including the unfair ones
- Even after you remove nationality, hidden stand-ins bring it back: surname, first language, neighbourhood, place of birth
- A model told only to "reduce vacancy" can quietly discriminate while looking efficient on a dashboard
- Studies in Europe and the US keep finding housing algorithms that hurt protected groups even when the obvious fields are removed
How aptari's Match Score avoids this
- It is a clear set of rules, not a model trained on old decisions, so there is no past bias to copy
- It scores only four areas: income, documents, listing fit and references
- Protected characteristics · nationality, religion, family status, age, gender, disability, sexual orientation, pregnancy · are never available to the score at all
- No name-based judgements, and location is used only to compare rents, never as a signal about the person
- Every score comes with the reasons behind it, so any applicant can see why they got their number
- The full record is available to the applicant if they ask
What no system can eliminate
No tool can stop an agency from adding unfair criteria of its own on top of the score. That is why the human decision is also recorded and the list of things the score never uses is published openly. Agencies on aptari agree to terms that require fair, lawful tenant selection.
See is AI tenant scoring legal and discrimination in tenant selection.