Skip to main content
In partnership with University of Strathclyde Masters of Engineering students

Hate speech attacks an individual or a specific group based on attributes such as sexual orientation, gender, religion, disability, colour, or country of origin. Some countries consider hate speech to be a crime, because it promotes discrimination, intimidation, and violence toward the group or individual being targeted.

Hate speech has been a hot topic of debate. Some people argue that any attempt to restrict someone’s expression of ideas amounts to a violation on his or her freedom of speech. Whilst others counter that hate speech does nothing but fuel the flames of violence and brutality.

What if Artificial Intelligence (AI) could detect and remove certain types hate speech depending on the level and severity of it Social media has become a mainstream medium for propagating hate, discrimination, racial abuse etc. targeting specific parts of society. This is being done mainly through text, audio, and video contents. Currently there is not any affective measures in place to automatically detect and block hate speech especially those in video and audio form.

As of 2020, the average daily social media usage of internet users worldwide amounted to 145 minutes per day, up from 142 minutes in the previous year (Statista, 2020). The current trends suggest, the youth aged between 12 and 18 are greatly influenced and shaped through the content they consume from all types of digital and social media channels. have been associated with Professor Durrani and 5 of the top University of Strathclyde Master of Engineering Students to work on detecting hate speech using causal AI. We helped frame the problem statement and research guidance for students to excel and protect people and society. The Strathclyde team has showcased their strategy, plan, and AI methodologies that is extremely innovative and should rapidly and proactively, detect various types of hate speech content to stop society from being negatively influenced.

The system needs research in the ‘Sound’ and ‘Video’ waves, to have intelligence inbuilt to create context and unravel deeper understanding of language, so it can detect more subtle and complex meanings. The tool should detect the emotional tone of a speaker and therefore analyse the most
accurate detection of Hate Speech. This will need research to create specific pattern recognition methodology, understand content holistically and detect hate speech. team will further collaborate with Strathclyde team,
upon their project completion and help them achieve this significant milestone.

We look forward to seeing the outcome of the Hate Speech detection model soon.

A diagram of how the proposed hate speech detection technology works: