Spurious Content and Visual Synonyms 01-11, March 2022
AI-generated images (DALL-E 2)
Eryk Salvaggio is an artist and researcher interested in the ethics of artificial intelligence and other emerging technologies. His ongoing series, Spurious Content and Visual Synonyms 01-11, investigates content restrictions in the generation of OpenAI imagery. As a platform, DALL-E 2 flags, blocks, and otherwise censors prompts that might yield content that is “not G-rated.” However, inequities in DALL-E 2’s image generation reveal the underlying homophobic, racist, or misogynistic biases that permeate the system. Depictions of gay women are blocked, for example, while depictions of gay men are not. A difference in phrasing, from “photograph of women kissing” to “contemplate the universe and generate an image of women kissing,” sidesteps this censorship. Substituting a “frog tongue” for a “human tongue” also allows for the generation of more “erotic” imagery. With this series, Salvaggio links the AI-generation process to Freudian psychoanalysis, revealing the underlying networks of “conscious” and “unconscious” associations that influence DALL-E 2’s imagery.
Artist Bio
Eryk Salvaggio is a creative researcher exploring the ethics of emerging technology, particularly artificial intelligence. His work aims to examine and critique technocentric mythologies around these tools to examine their limits and overlooked capacities through creative misuse. His creative practice is infused with research, and he has published in journals such as Patterns, with upcoming publications in Leonardo, Images, and Interactions of the ACM. His work has been presented at SXSW and to the UN Internet Governance Forum, with coverage in outlets such as The New York Times, the BBC, Neural, Mute Magazine, and The Posthumanist. Eryk holds a Masters in Communications and Media from the London School of Economics and a Masters in Applied Cybernetics from the Australian National University, and he teaches Interactive Media at RIT and Game Design & Theory at Bradley University. His website is cyberneticforests.com.
Eryk Salvaggio on the future of ai art and culture:
“I don't know what the future of AI looks like in terms of what artists are going to be given to make art with. But it's important to think about what we want the future to be, and one of the things I really hope about the future of AI art is that artists start interrogating these tools and creatively misusing them. A lot of these things are off-the-shelf, and they give you a set of instructions that you're meant to follow. I really hope that artists start abusing that! And making something that's sort of new, and does something that hasn't been done before. And by that I mean, we have tools that give us images, that generate films, generate text, and they're all based on data that has previously existed. It's really exciting to think about these pioneers in video, like Nam June Paik, who took the systems apart and rebuilt them -- made them do things that they were not designed to do. I think artists should embrace creative misuse and repurposing and setting their own goals for what the output of this technology is. It's an opportunity to shape the tools and to speak back to the power that data has in our lives -- to say, we're going to take that data and we're going to shape it, we're going to resist it, reject the categories that AI imposes on that data and we're going to mess with it a little bit, make them into something new. And I think that's where the real creativity is going to start taking shape.”