Critics are speaking out against an AI-generated expansion of Leonardo da Vinci’s Mona Lisa painting, among others, which fills the gaps previously left to the imagination. An image of the recreated painting shows an expansive background to the original painting that shows a canvas filled with imagery of the Mona Lisa’s backdrop.
The recreation uses the Adobe Firefly app which provides a Generative Fill feature that is drastically improved from previous versions and replaces the so-called clutter in images, with ones that are more pleasing to the viewer or creator. It is now being used to “improve” da Vinci’s masterpiece and has been plastered across social media, particularly Twitter, for all to see.
Jill Murray, a narrative director for game design responded to the newly created Mona Lisa on Twitter, writing, “The core problem with the current AI pitch is the assumption that we should always need or want more when choosing exactly what we mean to say and show is central to both art and communication. What lies beyond the frame was not chosen; that was the point.”
Artificial intelligence has received backlash in recent months as it enters the creative spheres including stepping on the toes of writers and artists. AI has sparked lawsuits and won art contests, sometimes editing a former artist’s work without compensation or consent.
“All of this art is taken without the consent of these artists and the laws that exist are not really protecting them,” Ron Cheng, a Yale Visual Arts Collective board member, told the Yale Daily News. “I think that there are enough artists out there where there shouldn’t really be a need to make AI do that.”
An image from the popular film Reservoir Dogs was also recreated with Generative Fillers used to fill in the space beyond the frame, and another used the feature to zoom out of the frames featuring close-ups of the film The Good, the Bad and the Ugly. Both received intense backlash, and film connoisseurs criticised the zoom-out change, saying those scenes were created deliberately and with great skill.
“Oh my god you picked the movie where a fundamental part of its style is that reality does not exist outside the frame,” one person wrote on Twitter. “Ebert wrote a whole thing about it. You literally couldn’t have picked a more anti-art use case.”
Others used the Adobe Firefly feature to use Generative Fill on Vincent van Gogh’s “Starry Night,” Michelangelo’s “Creation of Adam,” and Sandro Botticelli’s “The Birth of Venus.” These paintings are not the first to be affected by the AI “art” trend, albeit, in a different way. It comes only a few months after the Mauritshuis Museum in the Netherlands was criticised for choosing an AI-generated replica of Johannes Vermeer’s Girl with a Pearl Earring painting in a global competition.
One person commented on the museum’s Instagram post announcing the winner, “AI prompters are not artists and their AI-generated shell images should not be praised by the artistic community. Does no one at the museum understand how AI images are created? How the AI learns to create them? ***AI images are plagiarism***, how can you knowingly participate in that? So f*cking disappointing.”