Human-Oriented AI: Design Framework for Reaching Vulnerable Tenants
Principal Investigator: Viviana Bastidas Melo, University of Cambridge
Collaborator: Professor Jennifer Schooling, ARU
Council Partners: Cambridge City Council and South Cambridgeshire District Council
An AI-powered Early Warning System for Vulnerable Tenants system is being developed by the Cambridge City and South Cambridgeshire District Councils. Improving support for tenants is a priority for Cambridge City and South Cambridgeshire District Councils, but many of the systems involved do not speak to each other. Developing the technical solutions is not necessarily the ‘main’ challenge though. When applying AI to address this problem in the public sector, there is a need to consider what data is involved and whether it affects critical services and people’s lives or undermines public trust.
This project aims to develop a shared structure for governance, ethics, public value and data, to ensure these projects benefit the people concerned, rather than being led by the technology.
This Human-Oriented ethical framework for AI is designed across four layers; governance and management; public value creation; ethics and responsible innovation, and data, digital & infrastructure. It will incorporate the views of the groups of people concerned, including tenants and council staff, so that concerns, risks, ethical trade-offs and practical implications are documented and shape the design of digital innovations.
The framework will include diagrams and descriptions of the architecture of the Early Warning System, which show how each of the four layers are addressed, for example governance flows, public value outcomes, ethical considerations, data and infrastructure specifications. This gives the technical team a shared reference for the development and evaluation of solutions which are aligned with what the public and society needs. It ensures the initiatives are governed, legitimate, and responsible from the start, so that “reaching vulnerable tenants” is done in a way that city managers, tenants, and future partners can trust.