Impact case study database

Williams’ work provided a more accurate picture of a) the impacts of hate crime in Wales and b) hate speech online. The novel evidence provided, unavailable from other sources, underpinned the Welsh Government’s Framework for Action on Hate Crime and provided an effective way for UK police forces to respond to online hate. This allowed policymakers and practitioners to better identify, monitor, and prevent hate.

4.1 Creating a new framework for hate crime in Wales

Prior to Williams’ work, the Welsh Government did not have robust data on hate crime in Wales, as existing data sources – such as the Crime Survey for England and Wales – could not be reliably extrapolated to devolved administrations. Cardiff’s All Wales Hate Crime Project (AWHCP) was foundational to the Welsh Government’s hate crime policy through directly informing the national Framework for Action on Tackling Hate Crime [5.1] – the first of its kind in Wales. The Framework focuses on prevention, supporting victims and improving the multi-agency approach. Underpinned by Williams’ research, it cites the AWHCP more than any other research source (25 times). John Davies, the Welsh Government’s Head of Inclusion and Cohesion, confirmed that the Cardiff team “directly [fed] in robust data and insights from their research” and that the AWHCP “provided a much needed novel source of evidence” [5.2]. He cited three examples of evidence provided which were “essential to developing key parts of the Framework” [5.2]:

• Evidence on the impacts of hate crime on victims living in Wales across protected characteristics;

• Victims’ experiences of the criminal justice system;

• The emerging trend in online hate speech.

An evaluation carried out by the Welsh Government Audit Office (2015-16) stated “the Framework was…supported by robust evidence from the All Wales Hate Crime Project…. [it] helped…the Framework to focus upon areas which will have a significant impact” [5.3, p.1 ]. For example, under-reporting and the importance of community reporting systems were key findings of the AWHCP that fed into the Framework Delivery Plan. Welsh Government funding was allocated as a result to improve third-party (independent non-police) systems for victims to report hate crimes. Davies confirmed that “a major innovation of the Framework, supported by Professor Williams’ evidence that shows significant barriers exist to reporting hate crime in Wales, is the Wales Hate Crime Report and Support Centre” [5.2]. This Centre provides a flexible reporting mechanism, enhanced and integrated support for victims, and has a dedicated team of case workers. Davies highlighted an evaluation which found that the Centre contributed to an increase in hate crime reporting of 20%, despite victimisation falling by 28% in the same period [5.2].

4.2 Supporting delivery of policy actions and influencing practice

Williams is the only academic member of the All Wales Hate Crime Criminal Justice Board that takes forward the Framework’s recommendations. His ongoing work draws on HateLab research [3.3, 3.5, 3.6] to deliver the following actions from the Framework Delivery Plan (2015-6) [5.4]:

• Pilot a Community Tension Detection System to assess community feeling on social media during high periods of tension (the HateLab Dashboard);

• Conduct research on: i) the possible link between online hate speech and offline hate crimes and ii) identifying and modelling the spread of anti-Semitic content on Twitter.

In addition to informing policy, All Wales Hate Crime Project findings “have influenced practice and public engagement via Welsh Government outputs for end-users” [5.2]. Outputs developed using the findings include:

• Online Hate Speech Guides for use in schools, as well as by youth groups and probation services; these have been accessed 6,120 times [5.5a and 5.5b];

• Campaign materials for National Hate Crime Awareness Weeks (2013-2017) featuring films of victims of hate crime surveyed by the AWHCP speaking of their experiences, overlaid with highlights of the research findings [5.6];

• A Tackling Hate Crime Toolkit, which was distributed to all Housing Associations in Wales, has been used by a range of practitioners (including housing officers responsible for dealing with hate crime, managers, and contractors) to support training on estates with high levels of hate crime. The toolkit directly quotes issues highlighted in the AWHCP and also quotes a key project recommendation: “Social housing providers need to ensure that hate crime offenders are dealt with quickly and effectively and policies do not result in processes that ‘manage’ the victim rather than deal with the perpetrators (All Wales Hate Crime Research Project)” [5.7, p.12 ].

At UK level, the project has been widely quoted in national policy documents, including ‘Action Against Hate: The UK Government’s Plan for Tackling Hate Crime’ (2016) [5.8a]. Burnap was also invited, as a HateLab representative, to the Home Affairs Select Committee’s inquiry on Hate Crime and its Violent Consequences in 2016, set-up in response to the murder of Jo Cox MP and rising levels of hate speech and crimes. HateLab evidence was cited in the committee’s report showing that online hate speech could be detected at scale and in real-time with the ML developed at Cardiff [3.2 - 3.5]. The inquiry’s subsequent criticisms of social media companies for not using such ML to counter the spread of hate influenced the drafting of the White Paper on Online Harms [5.8b].

4.3 Tackling online hate speech for the UK’s National Cyber Hate Crime Hub

Police statistics on online hate significantly underestimate the extent of the problem as most incidents go unreported. The HateLab Dashboard, by monitoring in real time and not having to wait for victims to file a report, reflects a direct observation of prevalence. It was designed to facilitate near-real time responses to cyberhate, including targeting counter-speech at repeat perpetrators and pre-empting outbreaks of hate crime on the streets. It is currently used by Welsh Government Community Cohesion Coordinators, the Wales Extremism and Counter Terrorism Unit, and a London-based LGBTQ anti-violence charity [5.9].

The Dashboard has also been integrated into the UK-wide National Cyber Hate Crime Hub (2019), run by the National Police Chiefs’ Council. The Hub, the point of contact for all victims of online hate crime, uses the Dashboard to produce intelligence reports for police, senior civil servants and MPs. Paul Giannasi, Director of the Hub, is responsible for coordinating policy and operational activity across all Government departments and criminal justice agencies in their response to hate crime. He confirmed that the new data and insights provided by the research “transformed the way my team of police officers and civilian staff at the National Cyber Hate Crime Hub monitor and tackle online hate crime” and that the Dashboard has “fundamentally changed the way we monitor the spread of hate speech during national events” [5.10].

Prior to the Dashboard, Hub staff did not have an adequate way of tracking and responding to hate speech around events, such as the Brexit referendum vote. Giannasi stated: “during live operations, we were quickly inundated with irrelevant information and failed to capture hate speech in a systematic and reliable way” but the technological solutions provided by HateLab “ensures the Hub can monitor the production and spread of hate speech around events in a robust and reliable way” [5.10].

Dashboard data on the spread of online hate speech allows the police to better understand the dynamics of propagation, leading to improved response times, better support for victims and more effective allocation of resources [5.10]. Most recently, “the Dashboard has allowed us to gain valuable insight into how groups and individuals have used the coronavirus outbreak to stir up hatred online” [5.10].

Giannasi estimated that HateLab has resulted in economic savings of ~£500,000 via the Dashboard, cloud and data services and an implementation evaluation [5.10].