Artificial Intelligence and the Hate Speech Problem

In the present time, the freedom of speech and expression has been placed under a lot of scrutinies. The freedom to express one’s thoughts and opinions is truly a powerful thing. Social media is widely used to express opinions, ideas and beliefs. But with great power comes great responsibility. If every section of the public is allowed to have an opinion and express it on public forums like Facebook, Twitter, Instagram etc, then one has to be prepared that a lot many beliefs may be rather unpopular and hateful. Social media has become a breeding ground for trolls.

There are so many instances of celebrities leaving social media websites because of the massive hate they’ve received. But celebrities aren’t the only ones. Studies conducted by the likes of the Pew Research Center have shed light on how big an impact hate speech has on the common man as well.

Trolls and Hateful Comments Have Ruined Social Media

‘Saturday Night Live’ star Pete Davidson quit social media in lieu of the massive amount of hate he received due to his whirlwind romance and engagement to pop singer, Ariana Grande.


“No there’s nothing wrong. No nothing happened. No there’s nothing cryptic about anything. I just don’t wanna be on Instagram anymore. Or on any social media platform. The internet is an evil place and it doesn’t make me feel good.”

Pete Davidson on Instagram.

Singer Zayn Malik, back when he was with One Direction also deleted his Twitter account for the same reason. Although, he did return a short while later.

“The reason I don’t tweet as much as I use to is because I’m sick of all the useless opinions and hate that I get daily, goodbye twitter. My fans that have something nice to say can tweet me on the one direction account, x.”

Zayn Malik on Twitter.

It’s not just Hollywood celebrities that have faced hate. Indian celebrities have also quit social media for the same reasons. Actor Imran Khan quit Twitter after receiving a slew of negative comments. A lot of celebrities have also commented that they’d rather not join social media because of the negativity and hostility present.

“I feel that if you have a voice and you express something, you are often misunderstood. If you post a picture or comment, it is exaggerated 10 times over and I don’t want to get into that confusion. I would rather connect with my fans through my films.”

Ranbir Kapoor on social media.

Trolls are a Small Piece of the Puzzle

Hateful comments and celebrity trolls are however a very small part of the growing problem. A newer concern is that of hate speeches and promoting violence.

The recent surge in hate speeches has created a big problem for social media giants such as Facebook, YouTube etc. Curbing these would mean curbing one’s freedom of expression and speech. So where should one draw the line and how? Through their terms of service, social media companies have time and again stressed on cracking down on content that may be considered to be offensive or discriminatory. YouTube was embroiled in an ugly controversy displaying offensive and discriminatory ad content. This resulted in a loss of millions in ad revenue with big names such as HSBC and Marks and Spencer pulling out.

Monitoring content is thus a huge problem and a very critical one at that. It’s of equal importance to the users and producers of content on social media applications as well as social media companies. But how should one monitor content? By using artificial intelligence.

How Artificial Intelligence Helps

Artificial Intelligence or AI may seem like a superhero and it truly is one. Time and again, AI has shown a wide range of applications and has proven to be a highly versatile and vital technology. Mark Zuckerberg has commented that Facebook is in the process of developing AI tools to curb the hate speech problem. Let’s explore this further. AI has three main components – the data, the processing of the data and consequent action from the information retrieved.

Google also has an AI language filter called Perspective and Twitter has acquired Smyte, an anti-abuse provider.

The idea is to build an AI model that is intelligent enough to perceive context and independently judge whether the content in question is hateful or discriminatory. Given the gigantic amount of data being posted online every day, AI is the fastest and efficient way to tackle the problem. Human intervention is just not possible. It is not a very scalable option. Analysing copious amounts of data manually would require a lot of time and labour which is not viable.

“I am optimistic that over a five-to-10-year period we will have AI tools that can get into some of the linguistic nuances of different types of content to be more accurate.”

Mark Zuckerberg, CEO Facebook

AI’s Language Problem


The first problem is to build a dataset. Detecting language contexts is most certainly not the easiest thing to do. A dialogue from a movie or a quote from a book may be considered to be hate speech when it’s actually not. To train an AI model effectively, the ability to distinguish context is really crucial.

It is also important to be able to classify the content or the text or the image in question. This is also a part of understanding the context. The type of language used is also important. Some users tend to use expletives even with non-hate speech content. The AI model should be able to clearly differentiate between the two.

Languages are more complex than we realise. When human beings communicate, it is using some form of common sense. A machine or AI model may not have that common sense. This is what makes understanding context so difficult and challenging.

While artificial intelligence is a big part of curbing and reducing the hate speech problem, it alone cannot suffice. However, its still a start!

Posts created 4

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts

Begin typing your search term above and press enter to search. Press ESC to cancel.

Back To Top