As artificial intelligence develops further and increases its capabilities, the technology has received tremendous backlash over the past year. One of the uses for AI has been widely criticized in particular – AI art. New programs like DALL-E 2 and Midjourney use AI to generate art from short word prompts, and artists and illustrators worry that these AI programs could replace them due to the efficiency and relative affordability of AI art.
Last year, at the Colorado State Fair, Jason Allen submitted a piece created using Midjourney and won the digital art contest. Although he did disclose that his art was AI-generated, the New York Times reports that many artists were still enraged that Allen won the competition and $300 prize without creating the art himself.
Additionally, Cosmopolitan used AI to create the cover of their magazine, advertising that it only took 20 seconds to make. The Economist magazine did something similar, using an AI-generated image as an inset on the cover of their artificial intelligence issue.
But artificial intelligence might not only have a negative impact in the art world. AI can be helpful in art restoration. Art historians were able to be assisted by AI in recoloring paintings by the artist Gustav Klimt, reports The Wall Street Journal. One of his triptychs was lost in a fire, leaving only black and white photos of it, but recently they were recolored by an AI. The AI used references of other Klimt paintings and what critics had written about the piece before it burned to guess what it had looked like. However, not all historians agree that the recoloring is accurate.
Another use of AI that has recently come under fire is text-generating programs like Chat GPT, created by the company Open AI. English teacher Dr. Jason Brooks has been experimenting with it. There was a lot of worry in the academic community about the chatbot being able to write essays for students, but the chatbot has many limitations, he explained. “It doesn’t do analysis, and it doesn’t have style or voice.” But Dr. Brooks does believe teachers will have to adapt to this new technology. “My goal is always to ask questions that AI can’t answer,” he said.
Dr. Brooks believes Chat GPT could produce useful structures or models to help students with essay writing, similar to how students are taught the “three-prong thesis” or given an outline. It could also “provide starting points for creative writing.” He thinks it could eventually be a useful tool in research because its natural language abilities make it easier to find research without worrying as much about manipulating search terms.
However, Brooks does worry about AI being abused and replicating hate language. And in fact, Chat GPT users have reported many instances of the text-generator producing racist statements, The Intercept reports.
Artificial intelligence replicating racism is not a new problem; in 2015 a Google Photos service designed to sort photos into categories based on what was pictured mistakenly labeled 80 photos of Black people as photos of gorillas. AI is predominantly being built by white men, and biases can be embedded into technology, New York Times Opinion contributor Cade Metz writes. AI is trained by analyzing huge amounts of data. If that data contains biases, it can absorb that.
One app that has received backlash for allegations of embedded racism is Lensa AI, a program created by the startup Prisma Labs. The app uses pictures of people to create personalized “AI avatars” of them, turning users into fairy princesses, astronauts, and anime and sci-fi characters. But users of color also noticed the app often ended up lightening their skin as well, seemingly replicating colorism, the Smithsonian reported.
With artificial intelligence continuously being developed and seeming to expand its capabilities every time, it remains to be seen how, if at all, it will be regulated, and what effects that could have on our relationship with technology.