Introduction

As local authorities face austerity but demands for children's social care services continue to increase, many public bodies are looking to machine learning to better target their services to children and families in need. This project seeks to answer fundamental ethical questions about the use of machine learning in children's social care in England, and produce actionable guidance and recommendations for the future.

The topic of children's social care and the use of machine learning in the field can be quite contentious. Decisions made in the field of children's social care can have implications for the lives of children and families for years and the use of digital technologies to inform such decisions is still not well-understood (how it happens at the moment, nor how it should happen). This project aims to bring clarity to the field, but the topic remains sensitive to many.

Explaining the science

The report resulting from this research will provide concrete guidance for responsible machine learning innovation in children's social care. The research will clarify which ethical values and principles are common to social work and machine learning and AI ethics to create a 'Commitment to care, collaboration, and understanding' - a living document that should bring all stakeholders together to deliberatively decide on, and shape the path forward for, machine learning in children's social care.

The project will critically interrogate how these ethical requirements can be achieved in the reality and practice of children's social care and will conclude with recommendations for the path forward for machine learning in children's social care.

To complete its ambitious task, the project uses a range of methodologies, including desk-based research. The work is also informed by a stakeholder roundtable with local authorities, researchers, policy-makers, social care and family rights interest groups, and industry representatives, as well as by a workshop with family members with lived experience of children's social care.

Project aims

This project seeks to answer fundamental ethical questions about the use of machine learning in children's social care in England, and produce actionable guidance and recommendations for the future. It will address the questions:

  • Should machine learning be used in children's social care?
  • Can it be done right?
  • What is to be done?

Applications

This work could offer much needed guidance for machine learning innovation in children's social care policy and practice. In addition to researchers and interest groups, this work can benefit data scientist and industry experts working in the field, as well as policy and decision-makers in local authorities and central government.

Organisers

Researchers and collaborators

Dr Lisa Holmes

Director of the Rees Centre & Associate Professor, University of Oxford