Memes as an Antidote to AI Hegemony

Artificial intelligence (AI) has recently gained great progress in many disciplines, from natural language processing to visual recognition. In all these vast possibilities for computation, however, it seems that AI struggles at reproducing one of the most common human forms of expression: “memes”. Memes happen to be too complex as digital artifacts and too deep in being an expressive part of cultural, personal, or collective experiences for any AI to understand–despite its wide access to data. The failure emanates from the symbolic, cultural, and ironic layers impregnated within memes, which are composed of deep social contexts, nuanced allusions, and quick changes in culture. Even though AI processes massive amounts of information, it has never been taught to recognize or piece together the interplay of these intricate human dynamics. The problem arises from the very nature of ambiguity that humor, cultural reference, and irony possess. In other words, even with millions of examples of memes, AI’s attempts at humor seem to be far from human-like in most cases. In fact, the aim of this paper is to demonstrate that memes serve as an effective counterbalance to the growing influence of AI, because they offer a valuable insight into the limitations of AI in comprehending the nuances of human behaviour and emotional contexts. A comparative analysis of human-generated and AI-generated memes has enabled the identification of cognitive and creative limitations that define the boundaries of AI’s meme-making capabilities. Scientific relevance of this research is positioned at the intersection of AI and digital culture because as AI technologies come to mediate online interactions more and more, memes can offer a unique form of resistance against them, emphasizing the irreplaceability of human umor.