Attachment
The Dead Internet Theory: Why Being Online Feels Empty
Why online spaces feel isolating and alienating despite constant activity.
Updated January 4, 2025 Reviewed by Kaja Perina
Key points
- Dead Internet Theory suggests much of online activity is generated by algorithms, not humans.
- Online interactions often feel hollow because they’re mediated by simulations, not genuine human connection.
- Peer-to-peer communication fosters deeper, more authentic connections than algorithm-driven social platforms.
- Human connection is vital for well-being, but it is increasingly eroded online.
Dead Internet Theory originated on the fringes of the internet as a mix of conspiracy and satire about the rise of automation and digital manipulation. The theory suggests that most activity online is generated by algorithms and bots, not by humans. In its most extreme form, the theory imagines only one human remaining online, interacting with a web entirely populated by automated programs.
Dead Internet Theory may be hyperbole, but it raises two important points. First, the internet is undeniably moving in this direction. Meta, for instance, has announced plans to create AI users to populate its social networks. Second, and more significantly, the theory taps into a broader unease about the emotional and social dynamics online.
The theory touches on an issue that is both real and profound: the growing concern, malaise and disillusionment with the way automation and algorithms dominate online spaces. Many users report a hollow, uncanny feeling during online interactions: like shouting into a crowded void and only hearing echoes.
Digital Solipsism
But what drives this sense of digital solipsism? It’s not just the presence of bots or automation, it’s also the lack of genuine connection and attachment to others. According to attachment theory, forming meaningful social bonds with others is a fundamental biological drive essential for both physical survival and psychological well-being. Without real human connection, our online experiences can feel empty and alienating, amplifying the sense of shouting into a void.
Humans are social creatures. When we communicate, we rely on unspoken social protocols—norms that govern how conversations flow, how we read emotion, and how we build trust. These protocols are ancient and deeply ingrained. Online communication, though, is governed by computer protocols: instructions designed to manage the flow of digital information. These digital rules govern the platforms we use and shape our interactions with them.
Communicating Directly with Peers
Digital protocols can facilitate communication, smoothing the path for direct, uncensored or unmediated communication. It’s possible to use the internet for direct and meaningful connections. There are many tools online that let people interact directly, through text, voice and video. Peer-to-peer protocols are designed to connect people directly without inserting additional content or automated curation. Whereas many social platforms divert our attention toward automated activity and advertising.
Many platforms employ complex rules and systems that mediate and monetize communication. These platforms distort social connections by adding extra steps and automated activities aimed at optimizing goals like capturing attention, generating advertising revenue, or collecting data.
Living in Simulated Networks and Simulacra
Dead Internet Theory helps explain a particular relationship with technology. The technology is not mandatory, but it’s freely available. When you post on social media, your content is first tested against a predictive model: a simulation of your friends and followers based on their past behavior. Only if this algorithmic simulacrum "approves" is this post shown on that person’s feed or timeline. Communication is mediated by these predictive models, which act as social gatekeepers.
Mediated communication through bots and algorithms may disrupt our Theory of Mind: our ability to understand others’ thoughts and intentions. Social media platforms prioritize engagement over meaningful interaction, promoting posts that trigger strong emotions based on shallow meaning to spark reactions like excitement, outrage or envy. But we first have to catch the attention of the simulated version of people in our network if we want real people to see and react to our posts.
People trying to game the algorithm are engaging with simplified digital simulations of others rather than the actual people they represent. This can distort Theory of Mind by shifting the focus from understanding real human thoughts, emotions and intentions to predicting and appealing to algorithmic proxies, stripping communication of context and emotional depth.
Social media interactions have drifted into the French philosopher Baudrillard's concept of simulacra, where people engage not with real individuals but with algorithmic representations of them. These simulations decide whether content is likely to match engagement targets and subsequently be displayed to real people, distorting behavior into a spectacle aimed at pleasing predictive systems rather than developing human relationships.
This sense of emptiness is compounded by context collapse, where platforms merge all audiences. Friends, family, colleagues, and strangers merge into a single undifferentiated audience, rather than remaining individuated personal connections. Instead of addressing a specific person, group or trusted peers, you’re speaking to an amorphous, algorithmically constructed “everyone” made up of dozens or thousands of bots.
Peer-to-Peer Communication is Human
The internet wasn’t always like this. Early online spaces like email threads, forums and IRC chats operated on direct, peer-to-peer interaction.
You can bypass much of the automated activity online by communicating with others directly. In peer-to-peer (P2P) communication, sending a direct message to a friend, video calling family, or chatting in a private group you connects you directly, in real time, with real people. The social cues are created by human agency and influenced more by social rules than algorithmic protocols.
Compare that to posting on a timeline, where you’re essentially shouting into the void of bots, hoping to get noticed by digital agents and algorithms. Or doomscrolling through an endless stream of posts handpicked by algorithms designed to maximize your attention, not your connection.
Peer-to-peer connection requires more deliberate cognitive and emotional effort that scrolling a timeline, but it’s where the messier, personal and dynamic social interactions exist. Research across many fields of psychology demonstrates the importance of personal connections. From attachment theory, to evolutionary psychology and self-determination theory all indicate human connection is essential to well-being.
Reclaiming the human elements of social connection does not mean abandoning technology, but it requires designing and using systems that prioritize human contact, emotion and decision making. It requires addressing complexity instead of glossing it over, and spending time to explore the deeper social dynamics behind the people
References
Ball, J. (2024). Is anyone out there? Prospect Magazine.
Baudrillard, J. (1994). Simulacra and simulation (S. F. Glaser, Trans.). University of Michigan Press.
Cowan, E.T., Schapiro, A.C., Dunsmoor, J.E. et al. (2021). Memory consolidation as an adaptive process. Psychonomic Bulletin & Review 28, 1796–1810 (2021). https://doi.org/10.3758/s13423-021-01978-x.
Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104(3), 268–294. https://doi.org/10.1037/0096-3445.104.3.268.
MacRae, I. (2024). Web of Value: Understanding blockchain and web3’s intersection of technology, psychology and business. Alexandria Books.
MacRae, I. (2024). The internet isn’t dead, it’s just more fun at the edges. Digital Frontier.
Stein, B.S., Littlefield, J., Bransford, J.D. et al. (1984). Elaboration and knowledge acquisition. Memory & Cognition 12, 522–529 https://doi.org/10.3758/BF03198315.