GOVTECH CHALLENGE SERIES
How to detect and monitor hate speech on the internet?
Currently, hate speech and hate crime in the public internet space is monitored through a keyword-based monitoring system.
The number of cases provided by the keyword-based monitoring system is unreasonably high;
The system is inefficient as a lot of the time is spent on reviewing the cases, even though most of them are “empty“, i.e. no actual hate speech is detected; In other words, this requires an unsustainable amount of human resources that the organisation cannot allocate.
At the same time, a high number of hate speech cases go unnoticed, offenders are undetected, and a sense of impunity prevails, which in turn leads to an increase in such speech in public space.
Ministry of Social Security and Labour & The Office of Inspector of Journalist Ethics are looking for an innovative tech solution that can identify and monitor hate speech on the internet as accurately as possible.
The solution would be able to semantically identify hate speech based on given keywords;
Simple and user-friendly user environment;
Each comment must be related to the primary source of information;
Possibility to process and manage data provided by the system, e.g. statistics;
Possibility for continuous system improvements (developing system versions).
We are ready to provide all available information related to the development of the solution - keywords, examples of hate speech;
In our monitoring process, we have an opportunity to propose the introduction of such monitoring system to the media;
The organization would purchase the services of a properly functioning monitoring tool.
You can find more detailed challenge description here (EN & LT):
If you have other questions about the process, participation, deadlines, or anything related to GovTech Challenge Series 2.0, please contact us by email email@example.com.
Ministry of Social Security and Labour & The Office of Inspector of Journalist Ethics
You can find video of the challenge presentation (in Lithuanian) below: