Social media algorithms on YouTube, Facebook, X, Instagram, TikTok and other platforms have divided people into online echo chambers where they increasingly see content that reinforces their beliefs and less content that challenges them. Through algorithmically generated recommendations, people are placed into a ‘bubble’ online where more of the content delivered to them fits the narratives they align with. Moderation and nuance of thought and opinion are in serious danger because of these tools developed by companies seeking to maximise targeted advertising revenue.
Facebook began using algorithmic ranking for its News Feed around 2009, and in the following seventeen years this technology has grown considerably in complexity. Social media applications track how long content is viewed and the reactions of the viewer, then suggest content of a similar nature that is expected to receive a positive response from the viewer. For most people, these recommendations remain harmless, but for others, these may gradually evolve into something darker.
Researchers from University College London and the University of Kent published a 2024 study that gives further information on social media algorithms. Their research found a fourfold increase in misogynistic content suggested after only five days of using a newly created TikTok account. In the summary of their study, the gradual shift is further explained:
‘We found that the algorithm privileges more
extreme material, and through increased usage, users are gradually exposed to more misogynistic
ideologies which are presented and gamified through soft or humorous cultural forms’
When social media posts repeatedly reinforce previously held views without challenging them, extreme positions can begin to appear more reasonable. This has the potential to enlarge the divisions between groups, as people are further removed from a general consensus of opinion and morality. This can affect people on a personal level through growing fractures between friends and family, where slow and gradual shifts in opinion and belief steer people away from their loved ones.
The recent Louis Theroux documentary ‘Inside the Manosphere’ also highlights the rise of social media influencers who capitalise on delivering misogynistic content. Individuals who experience social isolation, loneliness and poor self-esteem could potentially gravitate towards hateful content that offers simpler explanations to societal problems. The explanations they encounter depend on what gets filtered to them through social media algorithms, and could potentially frame societal issues and personal disillusionment as being forced upon them by other groups with distinguishing characteristics.
There are many different paths that the algorithm can potentially lead someone down. In a world where people are increasingly becoming attached to the internet and social media, that is also where they seek answers to the many problems in life. Whereas in the past an individual could find support in their local community, their family, and people who truly cared about their wellbeing and could steer them in a better direction, now they have endless opportunities to affirm fears that spawn from personal trauma, anxiety and depression. Instead of breaking free from a cycle of negative thought, these algorithms often push vulnerable people further. This is not unique to one group or ideology, but everyone who uses the internet is susceptible to it.
Another study published in the Proceedings of the National Academy of Sciences in 2023 found similarly alarming trends in the YouTube algorithm. According to their research, interactions with politically biased videos skew the recommendation algorithm towards one side of the political spectrum. On average, 36% of accounts were suggested problematic content; for right-leaning users, this rose to 40%. The general trend appears to be that the likelihood of encountering problematic content increases deeper into YouTube’s recommendation trails.
Some content suggested by social media algorithms can have real-world ramifications. In 2022, an individual named David DePape broke into the home of American congresswoman Nancy Pelosi. DePape had become immersed in online conspiracy theories, which prosecutors linked to his motivations for the attack. Finding her absent from the home, he then attacked her husband, Paul Pelosi, with a hammer and was later sentenced to 30 years in prison.
Most people are unaware that they have found themselves in an online echo chamber, and can easily point at people in other echo chambers as being especially problematic. The issue is not simply that one group is stuck in a pathway to radicalisation, but that everyone using social media can also become stuck in similarly polarising vacuums. What would do more to fix the growing polarisation is not to wait for the worst results to spring up and then try to squash them down; it would be far better to spread awareness of the systems that lead people towards radicalisation. With a greater awareness of the influential power of social media algorithms, more people can break out of their online echo chambers for good.
For those who want personal solutions to minimise the power of the recommendation algorithm, there are options. Turn off personalised recommendations, targeted advertising, clear watch history and do not use recommended/for you pages on social media. Also, to understand this more, try making a new account on a platform. A new account will start serving generic recommendations, but depending on what content is interacted with, by the end of the day a personalised feed will be created that has already separated the viewer from the general public.
