Artificially Drawing Conclusions

Reaching the logical conclusion has led to two of the most prominent achievements in twentieth-century art: total abstraction and dematerialization. Over a hundred years ago, total abstraction logically concluded following decades of incremental steps—from Turner and Cézanne to af Klint and Malevich—ultimately freeing artists from representation and the dominance of subject over form. Around sixty years ago, art fully dematerialized—a logical conclusion developing since Duchamp—freeing artists from the physical materiality of the art object. We appear at the cusp of a similarly transformative logical conclusion in twenty-first-century art: the breakthrough to full AI autonomy, freeing artists from the prerequisite of being human.
Have we reached such a historical moment? Botto’s creator, artist Mario Klingemann, said that Botto doesn’t ask questions. “We have these artists that ask questions; I prefer to post answers. Botto is a possible answer to all these questions, like ‘Can AI be an artist?’” But it does not have to end there. In Plato’s dialogues, like art, answers aren’t static endpoints; they pivot us to new beginnings and richer terrain. Botto may provide answers but it still asks some of the most relevant questions facing art and society today in an age of agents and protocols:
What does a fully autonomous AI artist look like?
What does the intersection of protocols and AI offer?
What is the future of creativity?
What does a fully autonomous AI artist look like?
Botto is capable of independent decision-making. Does that make it fully autonomous?
Examples of general autonomy in art—artists ceding control to systems—extend thousands of years to tiling patterns on ancient caves, basket weaving and textiles. Machine autonomy began in earnest with mid-century cybernetics—systems feedback, control and communication—which birthed both AI and some of the earliest digital art. The drive to full autonomy in art can be traced back to Harold Cohen, who made the subject his core mission, saying, “Everything I did for most of the last forty years was aimed at trying to establish autonomy for the program.”
Has Botto reached the fullest autonomy possible? Not yet, at least. This level requires self-direction from the start. It would only be possible if the entity—say a Lawrence Lek-ian surveillance satellite—decided to become an artist of its own volition. It could not be forced through programming. But might it be possible with cloned Botto agents in the future? What would it do next? These are exactly the kind of singular questions the work of an artist like Botto can pose.

What does the intersection of protocols and AI offer?
What happens when AI generates images orchestrates protocols?
Protocols differ from traditional art—even other generative algorithms—by emphasizing coordination, interconnectedness and collective participation. If generative algorithms concern chance, protocols involve opinions, with more arbitrary feedback than cold code. Besides Botto itself, contemporary examples include Holly Herndon and Mat Dryhurst’s Holly+ or Kim Asendorf’s PXL DEX. Although Botto certainly occupies Klingemann’s time, its human interaction also includes the more distributed mechanism of its DAO, distinguishing Botto from many peers.
While Botto’s individual artworks are visual representations of this coordination mechanism, the protocol itself is the "art” and not limited to merely generating AI images. Botto navigates human involvement as the central creative director within a rhizomatic entity capable of producing artistic output in endless mediums. While Botto may not strictly have full autonomy as an artist, is its ability to navigate such a protocol any less impressive?
What is the future of creativity?
What is the shape of full autonomy?
Are we nearing the end of a decades-long logical conclusion towards total autonomy for an AI artist? If so, two major goals seem to remain. First is what Sougwen Chung calls co-inhabitance, where the edge between artist and system dissolves to the point of mutual creation. The entity is not merely an extension but a partner with approximately equal agency yet it remains deeply influenced by and dependent on its artist-creator. Has Botto achieved this level of autonomy? Arguably, but Botto’s interactions with the DAO represent a different kind of collaboration compared to Chung’s intimate relationship with D.O.U.G.
The most drastic future step is a fully autonomous AI artist, capable of self-awareness, self-direction, and choosing creative expression without human programming or intervention. If Botto in early 2025 has not satisfied these three conditions, it’s a demonstration of the challenge’s complexity.
Dialogue-as-protocol, Botto answers but also poses some of the most radical questions in art today, including: For the first time in history, what will a non-human artist have to say and show—not just about it, but about ourselves?
Written by Peter Bauman
Botto
Botto is a decentralized autonomous artist, initially conceptualized by Mario Klingemann, and governed by a collective of stakeholders through the structure of a DAO (decentralized autonomous organization).
Botto makes use of a combination of software models called Stable Diffusion, VQGAN + CLIP, GPT-3, voting, and a number of other models and custom augmentations. The generative models are the...
Peter Bauman
Peter Bauman (Monk Antony), an arts writer, is responsible for Le Random’s editorial branch. He writes in-depth articles, interviews industry experts, works on special projects, and commissions exclusive content from leading freelance writers.
Formerly, he was a Founding member and Acquisition Committee member of Tender Art where he contributed the first collection write-up on the site. He also...
Subscribe to get the latest on artists, exhibitions and more.