Introduction
Allocation of scarce resources is a recurring challenge for the public sector: something that emerges in areas as diverse as healthcare, disaster recovery, and social welfare. The complexity of these policy domains and the need for meeting multiple and sometimes conflicting criteria has led to increased focus on the use of algorithms in this type of decision.
With the advance of AI and the introduction of a vast set of algorithms, new avenues have arisen for automating the process of prioritizing people for receiving resources. Automation can potentially save time and resources; using algorithms for the problem also has the potential of mitigating human bias and increasing fairness in allocation.
Project aims
Beginning with the pressing issue of homelessness in the UK and collaborating with local councils that provide free housing to individuals and families, we are endeavouring to improve allocation decisions. Typically, the demand surpasses the available supply, necessitating a method for prioritizing individuals receiving resources. Through the predictive capabilities of machine learning algorithms, we aim to identify the most suitable resources for individuals, in order to improve their circumstances in a fair and ethical way. Furthermore, optimization techniques can enhance the overall well-being of candidates by setting distinct objectives in accordance with policymakers' preferences.
This system, initially designed for addressing homelessness, can be extended to encompass other domains of social care, such as patient scheduling in healthcare, organ transplantation, disaster response, and welfare programs.
Our ultimate objective is to develop a toolkit for various facets of social care, empowering policymakers to optimize allocation decisions by selecting specific objectives aligned with societal preferences.
Organisers
Dr Jonathan Bright
Head of AI for Public Services and Head of Online Safety
Vincent Straub
Research Assistant, AI for Public Services
John Francis
Data Scientist
Youmna Hashem
Researcher