Google Asks Artists To Defy Visual Cliches Of AI, With Striking Results

In Max Pinckers’ photo “Accidental Convergence,” six couriers in orange jackets, bulky delivery packs strapped to their backs, arrive at the same location in Brussels. An AI-driven navigation glitch brought them together, leaving a few looking around, likely puzzled at the sight of so many co-workers being directed to one place.

“The accidental convergence of these couriers reflects the occasional hiccups of AI-driven systems, reminding us of the possible dangers of relying solely on AI to shape urban dynamics, from the logistics of food delivery to the broader implications for labor and society,” Pinckers explains of the picture.

Pinckers is one of three artists commissioned by Google to expand the public conversation about AI’s evolving role in society through speculative photography that doesn’t necessarily rely on the technology itself. The goal of the “Alternative Images of AI” project, the company says, is to move beyond familiar visuals associated with artificial intelligence — robots, neural networks, computer chips and cables — and develop a richer, more layered visual vocabulary for what AI means, how it makes people feel and how it could impact society.

“The language we use — the metaphors, the mental models, the visuals — all shape the way we think about technology,” Mira Lane, senior director of Google’s interdisciplinary Technology & Society team, said in an interview. The more these elements tend toward abstraction, “the less people realize that this is a technology that is very differentiated,” she said. “It’s complex, it’s nuanced.”

Lane, an artist herself, founded Google’s Envisioning Studio, which showcases the possibilities of AI and other emerging technologies. The works created for “Alternative Images of AI,” 15 in all, are now on display at the International Center of Photography in New York City, where they will remain through Jan. 6.

Pinckers, an award-winning documentary photographer and an arts lecturer at Belgium’s University College Ghent, shot “Accidental Convergence” after witnessing a gaggle of couriers convene at a single spot in his Brussels neighborhood. He recreated the scene for Google’s project by staging it with with real people — no artificial intelligence involved — and then snapped photos.

“The situation made me think about how AI models have the power to direct and manipulate people’s movements,” Pinckers said in an interview, “and how this technology can be used to capitalize on human labor through often exploitative systems such as the gig economy. How roles have changed from technology serving humans to humans serving technology.”

AI Insights, From New York to Africa

Another participating artist, Charlie Engman, traveled more than 5,000 miles from his Brooklyn home for the project — to Ghana, which he had previously visited for his work with The Or Foundation, an NGO that works at the intersection of environmental justice, education and fashion.

On this trip to Africa, Engman asked more than 100 Ghanaians across the educational and economic spectrum to share their perceptions of AI, and their visions for its future. In Accra, the capital, he spoke with scientists exploring the clothing trade’s pollution effects, tech-savvy urban artists and designers and girls as young as 12 who earn meager wages transporting heavy bales of clothing atop their heads.

Engman, who has experimented extensively with AI-generated art, observes that the passionate and often heated discourse over the potential, perils and pitfalls of artificial intelligence predominantly reflects perspectives of those living in the Global North. “It was often missing a certain geographic and class perspective,” the art director of sustainable fashion brand Collina Strada said in an interview. “I was curious to talk to people who are not normally included in that dialogue.”

The artist’s images for the Google project emerged from his discussions on the ground in Ghana, which culminated in an AI wish list that often focused on tools that could be used to redistribute goods, wealth and knowledge more equitably and sustainably while adapting to cultural contexts and methods. Engman used photos he’d taken on previous trips to Ghana as prompts for the AI image generator Midjourney, then refined the output through further collaboration with the tool.

“It was a lot of trial and error,” said Engman, who opted to use AI for the Google project out of ethical concerns around consent and representation when portraying people. “I made a lot of ugly things that were too didactic, and I made a lot of squishy, poetic things that were a little too loose and held too little meaning.”

One image in the exhibit, titled “Kantamanto Registry,” physicalizes Ghanaian interviewees’ hope that AI could help inventory material goods at Accra’s Kantamanto, the world’s largest secondhand clothing market, which receives an estimated 15 million items of clothing weekly.

The picture depicts a bustling market of shoppers and vendors, with a jagged line reminiscent of those on graphs winding through the stalls. Data printouts hang from clotheslines, a representation of statistics made physical.

Another of Engman’s images, a surreal aquatic scene he calls “Algorithmic Altruism,” shows a group of people navigating through water on what appears to be a makeshift raft. The photo imagines an AI model for resource allocation that could ease the burden of people in drought-stricken areas by helping them find, purify and store drinking water. It’s just one potential use case for an AI model that would automatically know where and when natural resources are needed and identify solutions.

“We wanted to select a set of artists that had a distinct point of view and are experts in visually bridging the gap between reality and fiction,” said Lane, an artist herself.

The third artist participating in the project, Brooklyn-based photographer, filmmaker and musician Farah Al Qasimi, explores how communication with other living beings could transcend language and description to reach unseen realms of perception and intuition.

Her image “Aquarium” features a man gazing through glass at a cockatoo perched in a dreamy, vividly colored underwater world that he remains detached from. The photo originated with Al Qasimi’s longing to communicate with her sick dog. “I wanted to think of the possibilities of experiencing the world outside the limitations of a physical body,” the artist said in an interview.

ForbesGoogle Sends Artists Down AI Rabbit Hole To Reimagine ‘Alice In Wonderland’By Leslie Katz

Google has enlisted artists to interact with AI before. Earlier this year, Google Labs asked four visual artists to reimagine original Alice in Wonderland illustrations using artificial intelligence. Google Labs, which experiments with AI technologies, says it initiated the Infinite Wonderland project as part of an ongoing effort to involve artists in shaping AI tools.

AI And Art: A Complicated Collaboration

Reaction to AI among artists remains varied. Some are excited about its potential to steer them in weird and wonderful directions, while others express anger that companies are using their work to train AI datasets without credit or compensation. Many fear AI will eliminate jobs, and possibly undermine the very nature of creativity.

“Artists have a way of looking at technology profoundly differently than technologists do,” Lane said. “In some ways, they are an alarm bell for society.”