Pallas’s Cat There is a Picture in My Head

In WORKS

Eleonora Roaro, “Pallas’s Cat There Is a Picture in My Head”
Single-channel video | 2’00’’
1920×1080 HD | 16:9
Edition 3 + AP | 2023 
Vocals, lyrics and music: Eleonora Roaro
Sound engineer: bluEsForCE productions (William Novati)
Producer and additional vocals: In.Visible (Andrea Morsero)
Special thanks to Gabriela Galati

This project explores the potential of generative artificial intelligence (AI) in artistic creation, as well as its standardisation. It stems from a provocative question: can a catchy, potentially viral song be created with almost no skills in music composition, using AI and software? The resulting artwork, “Pallas’s Cat There is a Picture in My Head” (2023), is a pop song and video featuring an AI-generated musical base and lyrics, with vocals by artist Eleonora Roaro. The song explores the interaction between humans, animals and machines, and blurs the boundary between automatic creation and human intervention.
The inspiration for the project came from a viral reel on Instagram and TikTok featuring a Pallas’s cat or manul, a small wild cat from central Asia, in a zoo trying to keep its paws warm by putting them on its tail on a winter day. The video loops obsessively with a soothing effect, while the lyrics – written in the font commonly used for memes – appear on the screen like karaoke and ironically allude to the desire to appropriate online content, even if it could be dangerous.

Eleonora Roaro, “Pallas’s Cat There Is a Picture in My Head (Fur Pop Remix)”
Single-channel video | 3’00’’
1080×1920 HD | 9:16
Edition 3 + AP | 2024
AI-generated video clips: Siri Crespi, Alexandra Gripenhoftner From, Eleonora Roaro

The “Fur Pop Remix” inverts the making process employed in “Pallas’s Cat There is a Picture in My Head,” where the lyrics and the music were generated by AI while the video consisted of looped found footage. In this case, the audio adopts a traditional compositional technique, remixing, which does not create anything new but reassembles pre-existing elements. The video clips, on the other hand, are generated by employing frames taken from Instagram reels as prompts for the generative AI. This results in a montage in which each Pallas’s cat, in its transformation into the next, incorrectly morphs into a domestic cat, a tiger, or an imaginary feline. This occurs due to the AI’s comparative lack of specific knowledge of the Pallas’s cat or its movements, since other felines are far better known. Consequently, the AI processes the images based on the available data, leading to results that are at times amusing and at other times uncanny. The project demonstrates the current technical capabilities of generative AI, and reveals both its statistical and probabilistic nature as well as its manner of categorising which is distinct from that of humans.