The 'Dead Internet Theory' Gains Traction: Is the Internet Full of Bots?

The 'Dead Internet Theory,' which posits that most of the content on the internet is generated by bots, is gaining increasing traction. Even Sam Altman, CEO of OpenAI, has expressed concerns about the growing number of Twitter accounts run by LLMs (large language models).

The theory, popularized in 2021, claims that the internet has been 'hijacked by a small group of powerful people' and now operates mainly by and for bots. While in 2021 the idea seemed like a conspiracy, the evolution of artificial intelligence has made it seem more realistic.

According to Imperva, in 2024, 51% of internet traffic consists of bots, surpassing human traffic for the first time. The entry of artificial intelligence into content creation has exacerbated the problem, with the percentage of websites in Google's top search results containing AI-generated content increasing by 400% since the launch of ChatGPT.

This also has implications for content creators, as search engines provide AI-generated article summaries, reducing advertising revenue. Furthermore, if LLMs are trained on AI-generated content, there is a risk of high-quality content being depleted.

Adam Aleksic, a linguist, warns that the increasing use of AI-generated content could lead to a 'distorted, artificial version of public discourse,' as people begin to adopt peculiarities of LLM language.