Ethics and machine learning in children’s social care

22 February 2019

A few weeks ago we launched a call for local authority partners to work with us on a series of research projects relating to machine learning and children’s social care. Machine learning and other forms of predictive analytics are already being used in practice and so our aim with these projects is to explore – with an open mind – the possibilities and pitfalls of applying these techniques in a social care setting.

Understanding the ethics and acceptability (and not just whether it’s technically feasible) is central to understanding whether machine learning should be used in this context. With this in mind, we are hosting a discussion on the project on 27 March 2019 (6:30pm – 8.30pm) at King’s College London. We would like to have a full spectrum of views represented in this conversation to inform the research. Please register your interest in attending.

We are also commissioning an independent study on the ethics of applying machine learning in the setting of children’s social care (in general rather than specific to our current work). This review will be deliberately external as we hope that the reviewer(s) will bring the appropriate expertise and provide objective guidance for the sector.You can respond to our request for quote here: Machine Learning in children’s social care ethics review

The ethics review will run in parallel to the machine learning research itself. This is because reviewing the ethics of social workers using machine learning within their practice requires a careful weighing of potential benefits and risks. To assess what both of those are, you need an understanding of the quantified benefits – how much better, if at all, does it distinguish between children who need more support and those that don’t when compared with alternatives? – as well as the quantified risks – how much bias is there, if any? We envisage the review covering the ethics of the use of machine learning at different stages of the care journey and for use in different decisions.

This current machine learning research will be reviewed by a university ethics committee. While the Centre is still in its development stage, we are working to establish our own ethical oversight framework, to ensure the findings of any research we produce or co-produce in future protects the rights, safety, dignity and wellbeing of research participants.