Data ethics committee

The general aim of the committee is to move beyond standard committee models, which generally involve a project being scrutinised once and by the standing members of the group.

Specifically, this committee approach aims to do the following.

• Provide more effective and legitimate decision-making processes by asking the committee to review data projects at least three times and requiring it to act as a steering group.
• Not only think about the data use case in question, but be self-reflective and consider how it needs to make its recommendation. The committee focuses on the question, who else needs to be involved for this decision to be made legitimately? This allows for communities to be involved in any recommendation.
• Increase community involvement to expand decision making to those that are not often heard from. This increases the diversity of views that are applied to considering the ethical issues at hand. It also supports active, collaborative engagement with the public.

This initiative intends to support the development of legitimate recommendations, based on research conducted at the forefront of this field. By involving communities actively in making recommendations, the public become part of the police. This aids and improves the legitimacy of the recommendations.

Data ethics committee

Thames Valley Police (TVP) has developed the framework for a data ethics committee over several years. The framework is used to assess innovative uses of data, and was developed alongside academics from the Ethox Centre, University of Oxford.

This began with running a steering group to establish a terms of reference and ethical framework document, including ethical principles for consideration. This group included members of the violence reduction unit (VRU), where the committee originally sat, and experts in ethics.

Once a list of roles and people to fill those roles were identified, they were brought into the steering group before the first meeting of the committee in July 2021. While the committee originally sat within the VRU and a lot of the work began there, the committee has now moved into wider TVP.

The committee met every two months to:

• discuss various data use cases
• build on the approach in the terms of reference by supporting the identification of communities that are needed to be involved in making recommendations

This led to the development of a programme of engagement with young people around a use case presented to the committee.

Deliberative forum project

The engagement took place over two to three months, with different workshop sessions conducted with groups of students from local schools.

TVP ran two workshop sessions in the school environment. These aimed to introduce the students to thinking about police ethics more broadly, but also the specifics of data ethics.

Workshop one

Workshop one contained two activities.

The first activity gave examples of police practices in the news. One about a stop and search, and one about an intervention from the VRU that uses the public health approach to provide early interventions based on data. The students discussed the different approaches and what they liked and didn’t like about each.

The second activity then got them to think about different kinds of data that might exist about them and asked them to rank it as to how personal they thought it was.

Workshop two

Workshop two contained one extended activity looking to build on the previous session.

Students were asked to think practically about a problem. TVP created a fictitious story about someone’s life with different agencies involved that hold different information. The students were split into groups and acted as either the police, school or GP. They considered whether they wanted to share their information with any of the other tables, based on their ethical considerations. Once they made their decision, the sharing of any information was facilitated to see if it changed their views at all on their decisions.

Once this was completed, students were then given consequences for their choices and asked to provide a reason as to whether they were unhappy or happy with their decision-making.

To support these workshops, TVP had multiple table facilitators on hand. These included:

• one individual running things from the middle of the room and bringing conversations together
• three facilitators (one for each table) to support conversations among the groups

The facilitators had experience in the police and education as academic ethicists.

Real data use case

These events helped develop skills within the students to be critical about data use – which would be a fairly unfamiliar topic to them – in preparation for talking through a real data use case being discussed by the committee.

To discuss this use case, TVP brought 65 students from local schools together at the University of Oxford in November 2022 to build towards a recommendation. The students were split into eight different tables with a facilitator on each. TVP also had two people running things from the centre of the room when bringing views together, and a couple of people who were experts on the use case to present to the students.

The day was structured as a live thematic analysis, allowing the students to outline their ideas and gradually refine these to a recommendation as the day went on. This happened in live time on the day by pulling out themes from student discussions and refining them to a recommendation.

The students could recommend that there was either no intervention based on the data or that there was one. They were offered several options to choose between if they wanted to recommend an intervention.

They worked on tables together to come to a view, which then fed into a wider discussion to try to come to agreement as a room to provide a concrete recommendation. This has since been fed back into the committee to become part of their view and process for making recommendations.

No formal evaluation of the intervention has been conducted.

A full report on the young people's deliberative forum and workshop sessions can be read in the National Police Library. This report was written to try and capture the results of the deliberative event in as much detail as possible. This includes strengths, weaknesses, areas for improvement, and feedback and thoughts from facilitators.

The Ethox Centre has also written a blog post, providing high level explanation of the work. See Shaping data ethics: A young people’s deliberative forum (Ethox Centre, 2023).

The programme of work delivered with the students has heavily progressed the recognised need for wider scrutiny of innovative data practice. It has allowed TVP to develop the model for community involvement. This can be built upon and developed, with the hope that TVP can access more voices for future events to influence thinking. 

This model is for projects to be brought in the development stage and reviewed a minimum of three times by the committee. As part of this review process, the committee reflects on what it needs to do to make a legitimate recommendation – particularly who else needs to be involved or heard beyond the standing membership. TVP then uses a deliberative model. This is designed to engage people in such a way where their ideas are refined to an actionable recommendation by the body of people as a whole, which can then form part of the committee’s own recommendation.

TVP has also established a governance structure – including terms of reference and an ethical framework – to consider ethical issues surrounding data use. Issues could include bias in the data, for example.

What worked well for the data ethics committee

• Having an initial steering group before the first meeting – which considered the committee structure and model – supported buy-in for the model and ensured familiarity with each other before the first full meeting.
• Regular meetings between the independent chair and internal committee management ensured that more effective meetings can take place, with agreed structures for the meetings. These meetings have been held weekly.

Areas for improvement for the data ethics committee

• There is a need for a ‘thresholding’ process for what needs to come to the committee and what does not. This might be a lower-level ethical review process that establishes whether something needs to go for full committee review.
• There were challenges regarding attendance. TVP is resolving this through setting clearer expectations for membership and minimal attendance rules for a viable meeting. There will be expectations written into the terms of reference for members to make an effort to attend all meetings.
• There have been some challenges in terms of the number of projects going to the committee. Due to some delays in the projects going to the committee and the small number of people attending, there have been issues in having a consistent number of projects to ensure that meetings can go ahead.

What worked well for the deliberative forum project

• The workshops were necessary to ensure that those attending were comfortable and felt their opinions were valued before attending the main event. It also allowed the attendees to get used to speaking about their views on a complex subject, so they could give the best account of their views at the main event.
• The second workshop went particularly well. It supported the students to place themselves in the shoes of someone deciding whether to share data. Many of them realised that there were complex trade-offs that needed to be made. This became really useful for the main day, when they had to work together as a whole body of people to make a recommendation that would require negotiation of viewpoints to achieve.
• A detailed facilitator pack for the main day was produced. This provided clarity to facilitators on what they could do if conversation was slowing and a clear sense of what needed to be achieved by the end of each of the main sections on the day. It was noted by several of them that this supported the day effectively and ensured its success.
• It worked well splitting the day into the different parts and using each to refine the thoughts of the student group to a recommendation. This approach made it possible to manage a complex set of views and thoughts into a concrete recommendation that also managed the disagreements the students had.

Areas for improvement for the deliberative forum project

• The main day was very busy. The timings were thrown off very early on due to bad traffic causing all schools to be late. This was then hard to manage for the rest of the day, especially given the amount of time people need to be given to express their thoughts.
• There was some confusion about the main question for the main deliberative forum. Some time should be built into future events to make sure people understand it.
• For workshop one it may have been beneficial to provide students with a general understanding of stop and search powers. 
• For workshop two, TVP used consequences for student decision-making. There may have been better ways to run this to ensure the activity was more immersive towards the end. The consequences provided were less specific to the kinds of decision the groups may have made on their tables. TVP felt they could write these consequences in a more open-ended way in the future.