Early focus on data privacy for force machine learning tool

West Yorkshire Police is in the process of developing a Domestic Abuse Assessment Tool that explores the feasibility of using emerging supervised machine learning technology to enhance the completion of risk assessments at domestic abuse incidents by police officers and to help them make the best decisions. 

The force has just started working with the Information Commissioner’s Office (ICO) through a free service called a Regulatory Sandbox that supports organisations to create products and services which utilise personal data in innovative and safe ways.

Through the Domestic Abuse Assessment Tool, the force is trying to reduce the number of domestic abuse offences that are committed as well as trying to predict whether or not any serious harm offences may occur in the next 12 months. Through this approach, the force hopes to work with partner services at an early stage, putting in place a wrap around service to support families and prevent the situation from getting worse. 

‘This is a groundbreaking idea and way of working and West Yorkshire Police want to ensure that they comply with any data protection requirements and that they can identify any potential concerns at an early stage and rectify them in a transparent and fair way.’

Information Commissioner’s Office.

Drawing on ICO expertise and advice, participants can mitigate risks and embed ‘data protection by design’ in projects. The service has been running since 2020 and provides participants with increased confidence in the compliance of finished product or services.

Thames Valley Police previously used the Sandbox approach as part of a multi-agency project related to the Serious Violence Duty and in its end report the force described the Sandbox as a ‘critical friend’ and said it gave them the confidence to ‘design the project in a way which is privacy-friendly.’