Robots, says the critic and curator, do not make art: their task is to aestheticize goods. And the self-aware machine is very far away, perhaps unattainable

Domenico Quaranta is an art critic, curator and teacher with a profound interest in the ways in which technological changes in progress, and therefore also artificial intelligence, are influencing artistic practices.

For a month, Quaranta has been the artistic director of Spazio Vitale in Verona, a place of culture and discussion, of exhibitions and dialogue: a role that makes him even more suitable to discuss, with a secular approach, authors and authorship in the time of Artificial Intelligence strong>.

How is the daily life of creative people concretely changing with AI?

"It depends on 'creative', and what is meant by this word, which I personally don't like. Some artists are experimenting in interesting ways with artificial intelligence, adding it to their toolbox, sometimes collaborating with it in a process of co-creation.

Many professionals - from translators to programmers, from graphic designers to illustrators - instead feel threatened by its ability to replace the 'creative' in some of its specific functions: a threat that Furthermore, it only increases a transformation that has already been underway since tools with professional features, such as Photoshop or Audacity, became available on every desktop.

Another concern, much more tangible and concrete, depends on the fact that current AIs are trained on datasets built on unlimited scraping of content shared online, which allows them not only to create content, but also to create it imitating the specific style of a creator, even a living one".

What's inside this partially unpublished toolbox? Is it increasingly normal to use AI or are there more forms of rejection than we think?

"Talking about AI in 2023 essentially means talking about generators of text, sound, images and videos built on LLMs (Large Language Models), rendered available online in various forms, for free or paid use: ChatGPT, DALL-E, Midjourney, etc.

Using them is relatively simple and straightforward once you learn how to craft a prompt. Some use it to create a final product, others for research, inspiration, brainstorming.

The rejection is obviously there, and it can derive both from tiredness towards the aesthetics of these tools, which all in all are uniform and repetitive, and from the concerns highlighted below, which is also leading some to review your attitude to sharing online".

Will the art and creativity of the future be the art of refining the prompt?

"I don't think so. Knowing how to create a good prompt will probably be one of the skills that an artist will be able to add to the others already acquired, to be programmed to create a good primer for the colours".

How will the author's sign evolve? Will we feel the need to accentuate our authorship, precisely to eliminate any doubt that our work could be the result of a robot?

"Style, the authorial sign, is repetition, often forced by the needs of the market, which always asks for more of the same. There is no need to inconvenience Barthes and Foucault to question, today, the idea of the author.

From Duchamp onwards, the best artists have questioned the idea of style, building the continuity of their investigation on more impalpable and less reproducible factors".

Will we see a growing polarization between manual art, to claim the absence of artificial intelligence, and digital art, or will we go towards contamination?

"Let's not be fooled by the fact that what we call Artificial Intelligence is currently used above all to produce images and other cultural artefacts: AI is not an artistic tool.

It is an evolutionary phenomenon internal to the automation process of the means of production of late capitalism. Its primary purpose is not art, it is the aestheticization of goods.

He can use art, as he uses and has used other tools created for the same purpose: but when he does so, his is always an act of appropriation, a deviant and illegitimate use.

If there is polarization, it will not be between traditional arts and digital arts, but between artists who use AI in this way and creative industries that produce aestheticized goods."

Photoshop is introducing AI, but there was already, for example, Content Aware Fill to eliminate unwanted parts of photos: can we consider it a form of artificial intelligence? And how does it differ from the latest software that has been discussed for a year?

"The hype of the new tools based on LLMs, the effort to make them perceived as something radically different from what we are used to, makes us forget that artificial intelligence is already everywhere: in Gmail when warns you that you have forgotten the attachment; in the on-board computer of your car, when it signals that you are going off the road; in the Instagram filter which is activated only if you make a particular grimace; in the Content Aware Fill and in other Photoshop tools.

The sad truth underlying this cheerful lie is that there is very little that is new in current AI.

Despite the relatively modest progress made on the neural networks front, the spectacular results obtained by these tools are the consequence of other advances: more powerful and faster computers, an unprecedented ability to aggregate and analyze large quantities of data, exploitation of a invisible, global, poorly paid (or not paid at all) work.

Big data, gig economy and more powerful CPUs are the real factors of difference between the current phase of AI and the one that preceded it. And self-aware AI is stillfar, perhaps unattainable."