How social media platforms fuel extreme opinions and hate speech

When the internet was created in 1983 nobody could predict its effect on society. Many people saw it as a tool. Others were quicker to understand its potential to becoming something more. 

“The potential of what the internet is going to do to society, both good and bad, is unimaginable,” David Bowie told a sceptical interviewer while discussing his career in 1999. 

“I think we are on the cusp of something exhilarating and terrifying.” 

Hate speech on social media

That vision still rings true. On the one hand, platforms such as TikTok and Instagram have helped us to keep in contact with friends living on the other side of the globe as well as create new friendships with like-minded people. 

On the other, sites such as X (formerly Twitter) and Facebook facilitate the spread of controversy, conflict, and what is commonly termed ‘hate speech’. 

My research with Gaël Le Mens and Nikolas Schöll, of Pompeu Fabra University in Spain, concentrates on how the design of feedback on social media platforms affects the opinions expressed. 

Imagine you stumble across a post online. It is at the extreme end of what’s acceptable; nonetheless people seem to like it. You want to object, but what can you do?  

These platforms make it easy for consumers to signal their approval. Thumbs up! Heart! Retweet! But there’s no thumbs down. 

The feedback options on these platforms make it impossible to show or measure instant disapproval.  

Of course, anyone can leave a negative comment, but will the originator of the post wade through the noisy comments to see negative reactions? It’s a clunky, time-consuming way to gauge opinion.  

Encouraging digital abuse

Could this skew the content itself? Suspecting that this imbalance encourages users to express more extreme opinions, we set out to investigate. 

In so doing, we made several assumptions – which we backed up with real world experiment and analysis. 

We assumed that people tend to like content they agree with and dislike content they disagree with. 

We also assumed that the more ‘likes’ a post receives, the more likely an individual is to post similar material. And the more negative the feedback, the less likely a user might be to continue to post similarly extreme opinions.

Using our mathematical model, we compared two feedback scenarios on social media. In one, users could only ‘like’ a post. In another, they had the option to both ‘like’ and ‘dislike’. 

Our simulations showed that when negative reactions were made more visible, individuals were more reluctant to post such extreme opinions.  

The implications are clear – the design of some of the most popular social media platforms reward extreme opinions, in the absence of a ‘dislike’ option. 

What the originator sees at a glance is instant approval, rather than the vast number of people who objected, but had no easy way to express it. 

As a result, users are receiving the wrong message. What appears to be endorsement is in fact a lack of balance. 

The impact of hate on social media

And what does this imbalance mean for women or minority ethnic groups who receive a disproportionate level of abuse on social media platforms? 

Our work implies that social media feedback mechanisms encourage messages of hate and prejudice to remain and to be shared. Because objections to these extreme opinions are relatively obscured, both the creators and victims see a skewed version of reality. 

Authors of extreme opinions receive disproportionate support and start believing that society shares their belief. Victims are blinded to their potential allies and start treating social media as a more hostile place, devoid of support and balance. 

With this in mind, could social media companies restructure their feedback mechanisms to bring more balance? Is a simple button for negative feedback the best option? 

And if so, why do the likes of Facebook, Instagram, Twitter and TikTok neither provide a one-button option for a negative reaction, nor make it instantly visible if a user has lost followers or provoked some other negative reaction following a post?  

Why is there no 'dislike' button?

It’s complicated. Platforms assume that any negative reaction will discourage users – and this disengagement will hit advertising revenues. And nor does everyone like to give negative feedback. 

Our work shows that individuals are more likely to give a like than a dislike and leave a positive rather than a negative comment. And of the nearly 400 prolific social media users who took part in our study, 42 per cent indicated they wouldn’t want a dislike button, as it could be used to harass other users. 

That said, some social media platforms deviate from the simplicity of a single like button. 

Back in November 2021, YouTube removed the public counts of ‘thumbs down’ from all videos – meaning that although content creators could still see negative reaction to their videos, the public couldn’t. This was done, YouTube said, to prevent unfair targeting and harassment of individual creators. 

But removing viewers’ ability to see negative reactions has proven controversial – with many complaining they no longer had a means to judge whether a video was worth watching or not – a high negative count helped flag poor quality content and guided users in what videos they chose to view.  

Reddit is one of the few platforms with a single dislike button, although the downvote option is more commonly used to show if a post is irrelevant or violates community policies than to signal disapproval. 

The dangers of one-sided social media feedback

Our work also has wider implications for marketing departments and brand managers, suggesting that businesses should be cautious when using social media to gauge public opinion or to road test a product or campaign. 

A one-sided feedback design makes social media metrics an unreliable way to ‘read the room’. It can lead brands to be blindsided by a torrent of outrage when a campaign reaches a wider audience.  

Consider the launch of a risky ad campaign that touches upon sensitive topics. In late 2022, fashion house Balenciaga premiered a campaign involving young child models clutching teddy bears clad in bondage harnesses. 

The campaign triggered a furious and public backlash. Although the luxury brand swiftly apologised, the fallout from the ill-judged campaign was damaging and extensive. Although we don’t know what exactly led to this disaster, a contributing factor might have been over-reliance on social media metrics that misrepresented public opinion of risqué images.

What difference does negative feedback make?

There’s still much to learn about the complex world of feedback. The vast majority of social media users don’t interact at all with any content – instead they use feedback provided by others to judge the quality of content (as with YouTube) or form an opinion about what society thinks about an issue. 

Does the asymmetric feedback structure bias our view of the world? Would a simple negative reaction button bring more balance? And would users really disengage – or reconsider their extreme views – if the dislikes mount up? 

All we’ve shown to date is that feedback is imbalanced and that affects social media content. Users, from businesses to individuals, need to be aware - the disapproving hoards are simply invisible, unlike the ‘influencers’ who are rewarded with celebrity status for sharing their extreme views.

Read a version of this article at the Financial Times.

Further Reading:

Responsible AI - A new online tool to protect children from harmful content

Do online privacy regulations increase data-sharing?

Lessons from Tinder and Bumble on using boredom to drive employee engagement

Social media feedback and extreme opinion expression

Elizaveta Konovalova is Associate Professor of Behavioural Science. She teaches Foundations of Data Analysis for Management on the BSc International Management and BSc Management.

Learn how to lead technological innovation in the four-day Executive Education course Leading Digital Transformation.

For more articles on Digital Innovation and Entrepreneurship sign up to the Core Insights newsletter.