Introduction

Artificial intelligence (AI) initially began to make its presence felt on social media when Facebook, one of the biggest social networking platforms in the world, began widely using the advanced technology in 2013[1]. With nearly five billion people around the globe currently engaging with some type of social media[2], it seems inevitable that AI, particularly algorithms, would be used by platforms such as YouTube, Instagram, X, TikTok et al, to understand what their users like and dislike and keep them coming back for more.

Exploring the role of AI in social media and the implications for mental health, is therefore an important and pertinent area of study.

Most people accept that AI will become increasingly widespread in society in general, and on social media AI will continue to play a crucial role in the content we get to see and engage with, which means it is vital to examine the possible impact AI might have on a person’s wellbeing while using social media.

How social media uses AI

AI on social media platforms is primarily used to collect data and examine patterns. Data is valuable to advertisers and marketers, as well as content creators and a range of other sectors and industries, where it is important to understand and interpret what people think, feel and believe.

“There are many algorithms behind any large social media platform…One set of algorithms processes content. Another set of algorithms propagates it, that is, helps determine who sees what.”[3]

While using social media it is impossible not to interact with some type of artificial intelligence whether you want to or not.

The AI utilised by these apps, where human beings spend so much of their time, are primarily designed to keep us engaged but does this mean they also fuel our fear of missing out (FOMO) as well as other more debilitating anxieties and phobias? People with mental health challenges such as eating disorders, addictions, or depression, often turn to social media for advice or to find people who are experiencing the same issues. When this happens, the AI embedded in these platforms interprets this usage and sends similar content back to the user which can obviously lead to both positive and negative outcomes.

According to one study on problematic social media use (PSU)[4] which used a sample of 270 participants, “a vicious circle can ensue: Individuals check their Internet-communication applications more often because they do not want to miss out on something, and they want to be part of the online community. But as a consequence, they experience negative effects due to a compulsive-like or addictive use of communication applications.”

How AI interprets information

Once AI on social media begins gathering significant amounts of data on individuals and analysing it, the divergence from passive observation or voyeurism, to direct influence arguably becomes more likely.

So-called ‘recommendation algorithms’ are designed to know users better than they know themselves. This technology is what many social media platforms rely on most to keep their user numbers up, along with the daily level of engagement, which ties in with generating revenue from advertisers. Recommendation algorithms, therefore, feed (no pun intended) off what social media users tell them, and then they suggest it back to the user in a continuous cycle of supposed ‘quid pro quo’. The danger comes when this symbiotic relationship evolves from suggestion to persuasion and then possibly even further to manipulation.

This potential impact can be seen in the high-profile case of 14-year-old Molly Russell. Molly committed suicide after suffering from depression and subscribing to several online websites which, according to the coroner at her inquest, Andrew Walker, provided “access to images, video clips and text concerning or concerned with self-harm, suicide or that were otherwise negative or depressing in nature.” Mr Walker added: “The platform operated in such a way using algorithms as to result, in some circumstances, of binge periods of images, video clips and text some of which were selected and provided without Molly requesting them.” Mr Walker concluded that Molly died “from an act of self-harm whilst suffering from depression and the negative effects of on-line content”

Some social media channels have also been accused of using the information gathered by artificial intelligence to reinforce biases and create echo chambers which amplify people’s worst fears. Four peer reviewed studies published in Science and Nature looking into the influence of Facebook on the political polarisation of American voters, found that “algorithms are extremely influential in people’s on-platform experiences,” according to one of the leaders of the studies, Talia Jomini Stroud, director of the Center for Media Engagement at the University of Texas.

However, even when some changes were made to the way Facebook’s algorithms work, people’s political outlook did not significantly change.[5]

“Like prior control innovations, AI surveils, sorts, parses, assembles, and automates. And like prior forms of social surveillance and discipline, it weighs differently and more prejudicially on poor and minority populations. Far from being purely mechanistic, it is deeply, inescapably human.”[6]

Can AI manipulate behaviour?

“Social media platforms tend to amplify content that is extreme or sensational, as this is often the type of content that generates the most engagement and attention. This can create an environment where misinformation, conspiracy theories, and other harmful content spread rapidly.”[7] So-called ‘click bait’, which is defined by Merriam-Webster as “something (such as a headline) designed to make readers want to click on a hyperlink especially when the link leads to content of dubious value or interest”, is a classic example of manipulation which can have a negative impact on a person’s mental wellbeing.

This model or ‘tool’ relies on people being lured by sensationalised words or dramatic images to content on the Internet that might not even relate to the initial headline or photo that appealed to them. Users are encouraged to click on the story because clicks boost user metrics and potential advertising revenue, but the link might contain hurtful or harmful content or encourage the user to believe information that is not true.

Social media and mental health

In general, social media has been found to have a detrimental effect on the psychological health of its users[8] according to a systematic review of 16 studies conducted in 2020. This review concluded that “social media envy can affect the level of anxiety and depression in individuals”. Another systematic review of 20 studies published in 2022, indicated that “while social media can create a sense of community for the user, excessive and increased use of social media, particularly among those who are vulnerable, is correlated with depression and other mental health disorders.”[9]

When the AI technology used by social media platforms is added to the equation, with its potential to influence behaviour and possibly even manipulate users, it would seem inevitable that some short- or long-term impact on individual mental health and wellbeing cannot be avoided or ignored.

Conclusion

“Algorithms aren’t the whole picture: Just as important is the design of social media, platform processes, their incentive structures and, most critically, human-algorithm interactions.”[10]

AI is not perfect, and it probably never will be. In general, it is designed to gather and analyse information, and it does this via real-time interaction between individuals on social media platforms around the world. AI is therefore intrinsically limited by human beings themselves along with their biases and prejudices, as well as what we are willing to share about our lives and thoughts and whether those engagements are real and honest.

Some may argue, for instance, that if you are depressed or even suicidal, and you use TikTok to find a way to harm yourself, which leads the AI to conclude that this is the content you want to see, the system will only be doing its ‘job’ by sending you more and more information on that topic. It would be desirable if there was some mechanism built into AI algorithms to reinterpret that information, and instead send out helpline numbers or positive messages to someone who obviously needs help, even if they are not explicitly requesting it. Maybe that technology will exist soon from now.

In conclusion, AI on social media is having an effect on individual mental health, we know that from current research, but the same studies cannot conclusively say whether that influence is mostly good or bad. The implications are myriad; therefore, the potential impact of AI on mental health when using social media platforms needs further study.

“Moving forward, it is important to use AI tools in a responsible and ethical manner, with a focus on ensuring that AI-driven content is accurate, unbiased, and serves the best interests of users.”[11]

Karen Rollins  June 2025

References

[1] Brandom, R. How Artificial Intelligence is Shaping the Future of Facebook. The Verge. 2013. https://www.theverge.com/2013/12/17/5220914/how-artificial-intelligence-is-shaping-the-future-of-facebook - accessed August 2024.

[2] Wong, B. Forbes. Top Social Media Statistics and Trends of 2024. 2023. https://www.forbes.com/advisor/business/social-media-statistics/#source – accessed August 2024.

[3] Narayanan, A. Understanding Social Media Recommendation Algorithms. Knight First Amendment Institute at Columbia University. 2023. https://knightcolumbia.org/content/understanding-social-media-recommendation-algorithms - accessed August 2024.

[4] Wegmann, E, Oberst, U, Stodt, B et al. Online-specific fear of missing out and Internet-use expectancies contribute to symptoms of Internet-communication disorder, Addictive Behaviors Reports. 2017. https://www.sciencedirect.com/science/article/pii/S235285321730007X - accessed September 2024.

[5] Klepper, D. Deep dive into Meta’s algorithms shows that America’s political polarization has no easy fix. Associated Press. 2023. https://apnews.com/article/facebook-instagram-polarization-misinformation-social-media-f0628066301356d70ad2eda2551ed260 - accessed August 2024.

[6] Burrell, J and Fourcade, M. The Society of Algorithms. Annual Review of Sociology. 2021. https://www.annualreviews.org/content/journals/10.1146/annurev-soc-090820-020800 - accessed August 2024.

[7] Ienca, M. On artificial intelligence and manipulation. Springer Link. 2023. https://link.springer.com/article/10.1007/s11245-023-09940-3 - accessed August 2024.

[8] Karim, F, Oyewande, A, Abdalla, L et al. Social Media Use and Its Connection to Mental Health: A Systematic Review. National Library of Medicine. 2020. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7364393/ - accessed August 2024.

[9] Ulvi, O, Karamehic-Muratovic, A, Baghbanzadeh, M et al. Social Media Use and Mental Health: A Global Analysis. ResearchGate, 2022. https://www.researchgate.net/publication/357746877_Social_Media_Use_and_Mental_Health_A_Global_Analysis - accessed August 2024.

[10] Narayanan, A. Understanding Social Media Recommendation Algorithms. Knight First Amendment Institute at Columbia University. 2023. https://knightcolumbia.org/content/understanding-social-media-recommendation-algorithms - accessed August 2024.

[11] Mohamed, E. A. S., Osman, M. E. & Mohamed, B. A. The Impact of Artificial Intelligence on Social Media Content. Journal of Social Sciences. 2024. https://thescipub.com/abstract/10.3844/jssp.2024.12.16 - accessed August 2024.