Like other technologies that preceded it, artificial intelligence, commonly referred to as AI, is altering art forms that have existed for millennia. He would also be instrumental in creating art forms not yet imagined. In the future, there will be three or four types of intelligences creating art: human artists, humans and AI working in collaboration, machine learning models that do not require human initiation but which still lack sensitivity, and finally, in the event that technologists create truly sensitive objects. artificial intelligence, artificial general intelligence or AGI capable of self-creating art. In any case, the AI is unlikely to destroy the art, but it will certainly transform it – a process that has already begun. Humans are making choices about how this will happen, and they are making those choices right now.
At the end of the 19th century, the camera did not destroy painting, but it changed the world of visual arts. Freed from the pressure to portray reality convincingly, artists embarked on experiments with aesthetics and meaning. In the early 20th century, the readymade (the use of found and manufactured objects as art) did not destroy sculpture but freed some artists from narrow notions of authorship and craft. This freed the sculptors to focus on theory and concept.
Thus, one could say that the world has always asked existential questions about new technologies. Of course, it’s not entirely accurate to compare machine learning to a 19ecentury camera that doesn’t make decisions and can’t learn or learn to operate on its own. The merging of human practices with machine learning, as we move towards a future of AGI, points humanity to a visible horizon where human and mechanical intelligences intersect and interact. Or is it a precipice?
For his project, “ماه طلعت، Moon-faced”, Allahyari uses a series of carefully researched and chosen keywords with a multimodal AI model to generate a series of videos from the Qajar dynasty painting archives (1786 -1925). Through this collaboration, the machine program learns to paint new genderless portraits, with the aim of undoing and repairing a history of Westernization that ended over the course of non-binary gender representation in Persian visual culture. . (Morehshin Allahyari, ماه طلعت، MOON-FACED, 2022. Video and text courtesy of the artist.)
Today, visualization work can be quickly accomplished by AI, raising questions like what new forms of creativity it might unleash some artists and what their next experiments might reveal. Already, contemporary artists like Mario Klingman, Sasha Stiles, Lauren Lee McCarthy, Morehshin Allahyari, Carla Gannisand Connie Bakchi investigate this new potential by collaborating with machines; build their own training sets to demonstrate bias; and generally help the world talk about how deep learning and neural networks affect humans, both today and in a variety of speculative futures. Far from working to destroy art, these artists use AI to expand the field of creativity.
But it also deserves a critical examination of how humans will be used in return. Humans design the models and usually provide the training sets. It is already easy to see that the simple act of scraping the internet for training data encodes an explicit bias that puts marginalized communities at risk. It is imperative today to design the training process for Ais with intention, rather than closing our eyes and hoping for the best.
Moreover, the inputs to today’s machine learning regimes come from the same artists whose paid labor can be replaced by the models: illustrators, designers, writers, composers, translators.
That said, if artists and their supporters are dedicated to systems that advance culture rather than cannibalize it, AI can be an incredible opportunity for learning and collaboration. That’s a big if. AI training systems are far from set, and those creating AI programs are making crucial decisions right now.
Today, AI-generated images are created by models that retrieve existing images produced by creative work (along with their views and biases), as well as responses to those images, essentially creating variations on pre-existing images and stylistic choices. While this process provides the opportunity for collaboration and learning, it has also elicited a reasonable response from artists whose work has been used as training data without permission or compensation. Often the artists’ names – or the styles they helped create – are even used as prompts. And while the AI is regularly credited as a co-creator, the human creators whose work fueled the machine are often neither mentioned nor paid.
Fundamentally, it’s problematic to build a system that scrapes creative work and then expects that hardware to keep churning out to power the next generation of algorithms. In other words, if AI-powered creators are starved of inputs, there’s no reason to believe those outputs will continue to evolve. If companies use human labor to power machines, who designs the ethical framework, who applies it and who studies the impact? If the answer to any of these questions is who can make the most money in the shortest time, paying the least for creative work, then yes: the arts sector will suffer, both human and mechanically. As Jeanette Winterson points out in “Love(Lace) Actually”, from her 2021 collection, 12 bytes: how we got here. where we could go next, there is great potential for human-AI collaborations in novel writing software. But without Virginia Woolf, GPT-3 will not produce it.
“…if I type: Cat falls down the mine shaft and discovers a secret world of giant mice with computer skills, novel writer will help with all character components (probably lots of characters if he’s is about mice), and the twists and turns of the plot – and yay, I’ve written a novel (about mice). wake up one morning in Turkey as a woman, I probably won’t write Virginia Woolf’s. Orlando. Anyway, it’s done. »
If creative growth is to continue, now is the time to invest resources to equitably support and grow arts, education, libraries and culture – rather than taking the easy bait and allowing a few companies to now scrape existing art treasures for all. value, and maybe one day resell them to advertisers. Humanity has built a civilization of enormous wealth and abilities, even if they are not evenly distributed. Innovators and visionaries in the world of technology and art have the opportunity to choose the world they want. AI doesn’t make that choice, people do.
“As an artist who works with code,” writes Lauren Lee McCarthy, “I think in terms of scripts, both social and technical.” In LAUREN, McCarthy plays the role of a will-less personal assistant, mimicking the role of (usually female-voiced) programs like Alexa or Siri. Instead of shouting at an AI, participants shout (and sometimes, speak softly) for Lauren, who responds to requests without judgment. “I try to be better than an AI, understanding them as a person and acting in anticipation of their needs and wants.” Lauren Lee McCarthy, LAUREN2017. (Video: Testimonials LAUREN2017, directed by David Leonard.)
I don’t wring my hands at the thought of transformational change in art. In 50 years, art may be as unrecognizable to me today as Christo and Jean-Claude’s shrouded Reichstag would be to a European court painter. Yet this court painter would find much to recognize in Amy Sheridan’s painting of Michelle Obama. Today, society recognizes the merits of both works of art. One does not deny the other.
So those who create and support art should embrace this technology, yes, and learn from it, collaborate with it, and also make conscious decisions about the future. If the wealth generated by an AI workforce is used to create income security for those in need, for example, the world could see an explosion of creativity and diversity in art enhanced by technology. ‘IA. This type of explosion would see exciting and rewarding new forms of storytelling, community, and critique.
Today, as in the past, art is made for humans. People love it for its beauty, for asking questions and subverting the answers, for challenging society, for sounding the alarm, and sometimes, for simply making human beings feel seen. Even if humans merge with machines, it’s hard to see why they would stop creating or loving art.
In 2015, the internet was flooded with images created with Deep Dream, a program that used a neural network to find and enhance patterns in input images. Today, artists like Gretta Louw use technology with a heavy dose of meta-commentary. (Installation view of Gretta Louw’s On dark days we must dream in double time, 2021 Digital embroidery and digital printing on linen, and NFT .mp3 file via QR-Code. Courtesy of Honor Fraser Gallery, Los Angeles; photo by Jeff Mclane. )