Social media should no longer be treated as a ‘unified object’ and instead has splintered into three ‘separable categories’, according to researchers from The University of Amsterdam.
‘Towards a Post-Social Media Studies’, published in April, says: ‘What was previously treated as a unified object — “social media” — is fragmenting into analytically separable configurations that differ in how content is produced, how it circulates, and how users participate.’
Replacing it are three distinct formations. The first is ‘algorithmically governed broadcasting platforms’. These are the social media sites themselves. While the platforms were originally built around prioritising user generated content from people you followed — using likes, comments and shares to judge which posts attract the most engagement — the past few years has seen them shift away from this and instead display content based on passive metrics, such as watch time, completion rate, rewatching, and even momentary hesitation during scrolling. These latter metrics tend to reward content from professional accounts.
The explosion in the popularity of TikTok, with its focus on short-form videos and sophisticated algorithms, hastened the shift for the other platforms. Simultaneously, an increased scepticism around social media has made public posting less interesting, while generative AI has also meant that platforms are less reliant on individual creators, with AI-generated content now populating feeds.
The researchers say this has left social media platforms as broadcasting systems where ‘content flows from a relatively small and stratified set of producers…to large, algorithmically assembled audiences’.
Meta argued in a 2024 antitrust suit that it could not constitute a social media monopoly because it was no longer a social media company. Data submitted to the court showed that fewer than one in ten minutes on Instagram is spent viewing content from accounts followed by the user. Even the platforms see themselves as broadcasters now.
What separates them from traditional broadcasters is that each viewer’s experience is personalised. The system is a ‘continuous stream of content selected not because of who produced it, but because it is predicted to hold attention’. The researchers predict that the emerging end point is ‘content generated by models, curated by algorithms, and consumed by user-spectators whose attention supplies the data that refines the system’.
Those users looking for a more active experience are turning to ‘semi-private spheres and micro-communities’. These include private messaging groups within platforms like Instagram, TikTok and WhatsApp, subscription-based spaces like Substack, as well as semi-closed platforms around specific interests.
Finally, the researchers argue that AI chatbots can be understood as an ‘emerging media form’ with use surpassing active participation in social media. The study points out that 52% of the US population reported talking to AI chatbots in 2025, while only 35% of the population posted on Facebook, 20% on Instagram and 11% on X.
The shift to the latter two configurations, especially, marks an issue for media researchers because of the absence of public information and shared texts. The social media era rendered private communication public, but we are now moving into an age of total inversion. While the researchers don’t extend it to advertisers, it’s not hard to see how this makes understanding trends and reaching audiences more difficult — ultimately increasing reliance on the only group who have access to that data: the platforms.