Contestable Camera Cars: A Speculative Design Exploration of Public AI That Is Open and Responsive to Dispute

Contestability—the ability to dispute and challenge AI decisions—is essential for protecting human rights in local government AI systems. The paper argues that making these systems "responsive to dispute" helps ensure they respect citizens' autonomy and dignity, particularly when AI is used for automated decision-making that affects the public.

Current civic participation in AI governance faces significant representation problems. Interviews with 17 civil servants revealed that involving the public in contesting AI decisions is complicated by questions of who gets to participate and whose voices are actually heard, suggesting that simply opening systems to dispute isn't enough without addressing who has access to that process.

Public AI systems must align with existing democratic institutions rather than operate separately from them. The research found that for contestability to work in practice, cities need to integrate AI governance into established democratic structures and processes, rather than creating entirely new parallel systems for handling disputes.

Cities lack the organizational capacity to develop and operate responsible AI systems. Civil servants identified that local governments need to significantly expand their internal expertise and resources to properly implement contestable AI, suggesting that technical solutions alone cannot solve governance challenges without institutional support.

Speculative design can be a valuable tool for exploring responsible AI implementation. By creating a concept video of a "contestable camera car," the researchers were able to generate concrete discussions with practitioners about real-world barriers to contestability that might not emerge from abstract policy discussions.