Detect Then ACT (DeTACT)

“Taking Direct Action against Online Hate Speech by Turning Bystanders into Upstanders”

Featured Project.

Build a future.

168,000

Targeted population

4-6 years

October 2019 - August 2023

Support a classroom. Build a future.

ongoing0%

Background

In May 2016, the European Commission agreed with Facebook, Microsoft, Twitter and YouTube on a Code of Conduct on countering illegal hate speech online. 

In June 2017, the German government passed a Network Enforcement Act (NetzDG) to combat hate speech and fake news in social networks. 

In May 2020, U.N. Secretary-General Antonio Guterres appealed for an all-out effort to end hate speech globally.

Project

1

WHAT

DeTACT combines human activism with technology to monitor online hate speech and to create campaigns with positive responses. These campaigns will focus on promoting bystanders to stand up, fostering citizens to be better equipped to challenge harassment, racism or sexism on social media.

2

HOW

DeTACT uses new AI for detecting hate speech, to assist upstanders deflate toxic discussions in a self-regulatory approach. The project will also report illegal cases of extremism and radicalisation, following the EU’s Code of Conduct and local legislation.

3

WHY

Online hate speech has seen a disturbing growth in recent years, particularly in response to the migration crisis. Today, social media networks sometimes function as echo chambers, facilitating the anonymous expression of more radical views than in face-to-face interaction.

As a consequence, hate speech has been able to infiltrate various types of public discourse online. For example, in Germany there has been a rise in extremist propaganda on social media, openly inciting violence towards refugees. Europol has also observed that violent extremism is expanding, “partly fueled by fears of a perceived Islamization of society and [partly by] anxiety over migration”.

2

WHO

We are a consortium of universities, tech companies, NGO’s and citizens in Belgium, Germany and the Netherlands, supervised by a board of security, legal and ethics experts. The project is funded by the Rights, Equality and Citizenship Programme of the European Union (REC-RRAC-ONLINE-AG-2018, 850511, 2019-2021).

Over 400 upstanders are involved in the project, from the following initiatives:

Technology

WHAT IT DOES

We use AI to automatically discover hate speech. The AI is aware of thousands of offensive words and word combinations that constitute racism, sexism, aggression and dehumanisation. 


It is continually searching social media for publicly visible messages that contain such expressions. This technology is available to registered members of either the DeTACT Trainers or Upstanders Programme. 


The identity of the authors of hate speech messages is never revealed to the activists, however. Activists may choose to respond to a message, or it may be reported.

WHAT IT DOESN’T

If a message is not acted on, it will be forever forgotten by the AI after a period of two days. 

 

The AI does not have a memory of who posted what. In accordance with the General Data Protection  Regulation (GDPR), no personal data is collected, unless it violates law, or violates the Terms of Service of social media platforms, in which case it may be reported.