Local governments in the UK are responsible for all kinds of public services for their geographic areas and the people within them, including the care of the elderly, the provision of welfare to the vulnerable, and the day-to-day operation of community centres. In managing the procurement, financing, development, and delivery of this infrastructure, local authorities consequently collect a multitude of data about their communities and the people living within them – data that could, if harnessed with modern data science and algorithmic techniques, revolutionise the efficiency and optimality of public service delivery.
Our framework for AI-ready data has provided a strong baseline for assessing whether datasets are fit for the next step of the AI revolution. For councils, AI-ready data could enable them to algorithmically forecast budgets for social care, proactively prevent homelessness, or computationally optimise transport networks.
This report contains 10 case studies in which we utilise both interviews and desk research to understand local authorities’ conformance to the blueprint laid out by our framework. There are more commonalities than differences, with data across the UK sharing both strengths and weaknesses with regards to dataset properties, metadata availability, and infrastructural accessibility.
Importantly, each strength and each weakness arises from a decision, a choice made during the design, development, procurement, iteration, or upkeep of data. Such decisions are not easy to make, especially in the ever-evolving landscape of data and AI. We therefore hope that this research can provide a structured evidence base for leadership, especially in resource-constrained, under-pressure organisations like UK local authorities, to make sure they navigate decisions correctly and build data (in addition to products and services built on it) that is at the cutting edge, ready for AI.
