Scientific publication demands accuracy and reliability, including in visualizations. However, achieving the expected level of completeness is not always possible due to physical and linguistic constraints. Space and nano-objects exemplify the inherent difficulty of visualizing objects “as they are.” This study examines visual representations of Artificial Intelligence (AI) through an analysis of 272 images, including covers of scientific journals and books, as well as article illustrations. To date, the dominant visual tropes depict a human/android figure or a brain, often combined with references to technology (circuits, chips, binary code, etc.) against a blue background. These images represent not AI itself, but rather its projection into a speculative future – a somewhat incongruous approach for scientific communication. However, a recent trend has emerged challenging this spontaneously formed anthropomorphic imagery through more complex and ambiguous visuals. This study identifies a nascent trend depicting AI as a circle filled with diverse shapes, lines, and colors, conveying a sense of intricate richness. This imagery can be interpreted as representing a self-contained system encompassing information of diverse natures. Unlike the futuristic projections, this more deliberate approach visualizes AI as a phenomenon already existing within the contemporary socio-technological environment. Nevertheless, this conceptualization is not yet firmly established.