The Reporting Experiences of Survivors to Analyse in Real-Time (RESTART) project developed by Trilateral Research has won the title Best Use of Data for Not-For-Profit or Non-Commercial Purposes. The project addressed a critical challenge faced by NGOs: enabling efficient extraction of valuable data that has long been obscured due to resource constraints. This approach enabled one partner – Causeway – to navigate volumes of information, identify real-time trends, and enhance their ability to provide impactful support. It also allowed another partner, Aberystwyth University, to develop evidenced based policy recommendations for better supporting survivors of modern slavery.
Making a difference
Since the introduction of the National Referral Mechanism (NRM), the UK government’s dedicated framework for identifying and supporting survivors of modern slavery, debates have arisen over the effectiveness of current provisions in meeting survivor needs. These discussions coincide with a growing recognition of the necessity to incorporate survivor perspectives into our understanding of their recovery from the traumatic crime they experienced, and the essential support systems required for its facilitation.
The RESTART project employed an innovative blend of artificial intelligence systems, data analytics and subject matter expertise to analyse large datasets related to survivor support needs. Recognising the authority of survivors in comprehending their own needs and proposing solutions, project RESTART introduced a novel approach to integrating survivor voices in research and policy making, using cutting edge ethical AI.
Despite a suite of reforms being announced in 2014, the Home Office has been slow to implement key changes that respond to shortcomings in support of modern slavery survivors. Experts on Project RESTART recognised that the Home Office needed evidence so that it can adequately reform its current support mechanism. That evidence should come from the true experts – the survivors.
Recent research has sought to involve survivors in the identification of survivor support needs through the integration of advisory panels and peer researchers. Most prominently, survivors were involved in the development of a Modern Slavery Core Outcome Set (minimum standard set of outcomes that should be measured and reported when assessing the effectiveness of programmes that support survivors) as peer researchers, through interviews and as workshop participants.
Due to limited resources, such research engages with a small number of survivors, between 10-15. However, AI technology and data analytics capabilities enabled project RESTART to analyse datasets which detail the self-identified needs of a much wider group of survivors of modern slavery. Trilateral Research analysed almost 40,000 stories. The use of technology overcomes barriers to wider survivor engagement, such as the lack of accessibility, while mitigating some unintended consequences of interviews that can have traumatising effects on both the subject and researcher. This approach set out to utilise Natural Language Processing (NLP) in the analysis of complex data sets held by Causeway that would otherwise remain underutilised due to resource constraints.
In its first phase the project focused on utilising NLP to analyse the information captured in Causeway’s support worker case notes. Causeway, like many other NGOs, manages a significant volume of case notes that document service user support needs and the corresponding actions taken to address them. These notes contain rich and insightful survivor-centred data, which has remained largely untapped due to the sensitivity of the material and the resources required for analysing such extensive volumes of information. NLP presents an opportunity for streamlined extraction of relevant data in a swift and efficient manner, with the potential to identify emerging trends in real time.
Insights were extracted using NLP and data analytical methods, and subsequently categorised into the following key areas of support need:
- Financial
- Employment
- Social integration
- Safety
- Education
- Legal
- Physical health
- Mental health
- Accommodation
- Dependants
The project derived patterns and trends on the needs of survivors of modern slavery using 39,513 unique case notes, facilitated by a data sharing agreement. These notes documented interactions with 545 distinct service users who received support from Causeway over a period of more than five years. Based on this analysis the team developed evidence based, concrete recommendations to be taken up by policy makers.
The knowledge gained from RESTART and from working with other NGOs and public facing organisations has allowed Trilateral Research to develop Honeycomb – an ethical AI solution that helps understand exploitation at a local, national, and international level. For the first time, the public and third sectors have a solution that provides a data-driven, evidence-based approach to the understanding of, and response to, exploitation and violence, that can support policy change.
