An analysis of the modern ‘dead internet theory'
Gabriel Dearden · Aaltodoc (Aalto University) · 2026
Generative artificial intelligence (AI) has given rise to an increase in the creation and proliferation of synthetic content on the Internet and, as a result, the previously dismissed ‘dead internet theory’ (DIT) has experienced renewed relevance. This thesis aims to evaluate the extent of the influence of AI on online knowledge and authenticity and subsequently determine the possible risk posed by model collapse. To address these aims, this thesis conducted a literature review, combining academic studies and digital safety reports to measure the amount of online AI-generated content and analyse model collapse. The findings of this thesis identified that: 1) while the original DIT post is largely sensationalism, a modern, ‘leaner’ DIT has legitimacy due to a quantifiable increase in synthetic content and AI exploitation; 2) model collapse does pose a systemic risk, however it is not inevitable and can be comprehensively mitigated through the use of data accumulation and ‘real’ data; and 3) mitigations and solutions to the misuse of AI, inauthentic content, and model collapse are applicable and in development. The thesis concludes that generative AI has driven forward the creation and proliferation of synthetic content and information. Consequently, the establishment of effective and accessible mechanisms to alleviate the detrimental impacts of AI-generated content is essential to preserve ‘real,’ authentic, human contributions on the Internet.