Annual Report 2019 – Software Tools to Combat Hate

This article is an extract based on OHPI’s Annual Report for the 2019 Financial Year. Return to the Report.

Software Tools to Combat Hate

OHPI’s proprietary software programs are the tools we have developed to support our fight against online hate. The Fight Against Hate monitoring software and its related CSI-CHAT analysis tool have been designed to work together in providing empirical data for analysis & reporting.

The Fight Against Hate software has been under development since 2011. This tool was intended as a technological solution to the challenges of dealing with the emerging threat of hatred expressed online. It provides shadow reporting and independent data which can be utilised in formal reports to both government organisations and social media platforms.

The Fight Against Hate software was officially launched in 2014 and following its use to produce the world’s first major transparency report on antisemitism in social media for the Global Forum to Combat Antisemitism in 2015, the tool was formally endorsed by a resolution of the gathered experts.

A report into Islamophobia followed, sponsored in part by the Australian Federal Police and Islamic Council of Victoria. Two UNESCO reports in 2015 praised Fight Against Hate calling it one of the “innovative tools for keeping track of hate speech across social networks and how different companies regulate it”.

As a result of its success as a reporting tool, other countries and organisations began to request the use of Fight Against Hate. This led to a redevelopment of the software to increase its flexibility and reduce operational costs. Work was done to refine and expand the Fight Against Hate software and in turn, to develop a further tool, a prototype system to work as an analysis platform. This has been called CSI-CHAT (Crowd Sourced Intelligence – Cyber-Hate and Threats) and works in real time with the data from Fight Against Hate.

A number of teams of final-year Monash University students have been working with us on both the CSI-CHAT tool and the redeveloped Fight Against Hate tool. This work culminated in the production of a new version of the reporting software, which allows the tool to be embedded as a widget in websites or Facebook pages of organisations. The categories that can be reported are configurable for each organisation, as is all text within the system – allowing it to be translated and to run in a range of different languages or simply with different wording as appropriate. With different configurations, the tool can take reports of content ranging from antisemitism or Islamophobia through to foreign interference or incitement to violence. Details showing what has been reported through a particular organisation can be visible to organisational administers in real time. All data from across all organisations can be immediately accessed and analysed by users with the appropriate permissions in CSI-CHAT. Working together the tools empower individuals, community organisations, researchers and governments.

The latest work, occurring during this reporting period and completed in late 2019, introduced the “review” feature from the original software to the new version. This improves quality by allowing the system to assign an item one person reports for other people to review. Transferring this feature from the original software to the new version makes the software more complete. The “review” capacity is a key feature of the system’s quality control and anti-abuse measures. While CSI CHAT is still a prototype, Fight Against Hate version 2 is now operational and a version has been configured for use in Italian by a partner organisation and a project based on it will be launched in Italy in 2020. Early discussions are underway with other countries and organisations.

Share this important work: 

Please help us keep these tools online by making a donation to support this ground-breaking work.

This article is an extract based on OHPI’s Annual Report for the 2016 Financial Year. Return to the Report.