Peeking Inside the Schufa Blackbox: Explaining the German Housing Scoring System

Different users need different explanations: The study found that tenants and landlords have conflicting information needs when it comes to understanding housing credit scores, suggesting that a "one-size-fits-all" explanation approach is insufficient for responsible AI systems.

Context and role matter for transparency: Rather than simply making algorithms more transparent, effective explanations must be tailored to users' practical realities and how the scoring system directly affects their lives—a key principle for human-centered AI design.

Current systems lack role-specific design: The research on Germany's Schufa housing scoring system reveals that existing explanations don't adequately address the distinct concerns of different stakeholders, highlighting a gap between how AI systems are currently explained and what users actually need to understand.

User agency should shape AI explanations: The study emphasizes that responsible AI requires considering not just what information users receive, but how their specific position and power in a situation (tenant vs. landlord) influences what explanations would be most meaningful and fair to them.

Speculative design can improve AI accountability: By having students imagine user interfaces from different perspectives, the research demonstrates how design-thinking approaches can help identify blind spots in AI transparency efforts before systems are deployed to real users.