The Digital Zombie Apocalypse
The "Dead Internet Theory" is a well-known conspiracy theory that has been circulating for approximately a decade. It originated on platforms such as Reddit and 4chan between 2015 and 2017. Given the current state of the information era, where Artificial Intelligence makes some sort of 'groundbreaking' advancement every six months, I'd say it's time to give this theory a second look.
Let's first establish what the 'Dead Internet Theory' entails. The theory posits that a significant portion, if not the majority, of internet traffic is driven by bots and AI-generated content. Some more conspiratorial versions of the theory contend that this content is created by government agencies or other parties with malicious intent. This is achieved through algorithmic curation aimed at manipulating public perception.
For the purpose of this discussion, we will focus on the non-political aspects of the theory.
This theory gained significant traction in 2021, particularly after the release of OpenAI's ChatGPT. Publications such as "Dead Internet Theory: Most Of The Internet Is Fake" on the forum Agora Road's Macintosh Café, as well as an article titled "Maybe You Missed It, but the Internet 'Died' Five Years Ago" in The Atlantic, contributed to its prominence. The claims made by this theory suggest that bots and bot-generated content have replaced human activity on the internet.
But what empirical evidence supports these claims? Let us examine some peculiar behaviours observed on major internet platforms.
For instance, on Facebook, there is evidence of a surge in AI-generated art posts, such as the infamous 'Shrimp Jesus' phenomenon that occurred between 2021 and 2024. These posts received a deluge of likes and comments, often consisting of similar messages. According to the theory, this exemplifies bots generating AI content and collecting artificial engagement from other bots. Similar observations have been made with various posts featuring AIgenerated content, including images of impoverished children alongside artwork supposedly created by them.
Furthermore, when analyzing two provided visuals, it becomes evident that while the number of Facebook accounts has skyrocketed, the average time spent by individuals on the platformis barely thirty minutes. This has led many to believe that the platform is inundated with bot or 'spam' accounts.
Similarly, on Twitter, numerous accounts have been observed tweeting phrases like "I hate texting," followed by banal activities. Strangely, many of these posts garner thousands of likes, suggesting that they are generated and engaged with by bots.
Meanwhile, on YouTube, there exists a practice known as 'the inversion.' This involves the market for fake views, which are used to enhance a channel's credibility and manipulate the algorithm in order to reach a wider audience. The prevalence of these fake views has causedthe algorithm to misclassify them as genuine organic viewership. This example underscores the theory's premise that the internet is flooded with bot-generated content and lacks authentic engagement.
Furthermore, the platform 9GAG, traditionally used for sharing and viewing memes, has apparently been overrun by bots that post propaganda, repost existing media, and contribute nonsensical comments. Over the years, the website has witnessed a decline in quality, with accounts only a few months old flooding the platform with repetitive posts and comments. The ongoing conflicts in Ukraine and Palestine have exacerbated the situation, as war propaganda is now regularly posted and reposted, inundating the website's feed.
Even disregarding the aforementioned examples, countless instances supporting the theory can be found across various platforms with a simple Google search. According to the theory, the 'death' of the internet occurred in 2016-2017, when artificial content surpassed organic content in terms of volume. Reflecting on our own internet experiences, we are likely to recall instances of low-quality media becoming inexplicably popular or encountering frequent reposts.
Looking ahead, language models continue to improve, capable of generating text that appears organic. Image generators produce artwork more vivid than that created by actual artists, and video generation technology has also made remarkable advancements. It is worth considering the groundbreaking deepfake video in which Jordan Peele impersonated President Barack Obama, which occurred six years ago. Anticipating the future, it is conceivable that distinguishing genuine media and engagement from fake stuff is getting harder and harder now, and soon it could be comparable to finding a needle in a haystack. With that, all I'd like to say is - maybe it's time we start looking more carefully at the content we see online. content will become increasingly challenging, akin to finding a needle in a haystack.
In conclusion, it may be prudent to approach the content we encounter online with greater discernment. The Dead Internet Theory, a well-known and debated conspiracy theory, contends that a substantial portion of internet traffic is driven by AI bots and generated content. While it is essential to critically evaluate the evidence supporting these claims, the observations made on major platforms offer notable insights into the functioning of the contemporary internet.
Neville Spencer James, Chief Technical Officer, Analytical head, Lila Creations.