AI art tool MidJourney has all the answers to ‘what if’
Inspired by NASA’s recently released images of the universe, the first hint I saw at the research lab MidJourney’s Artificial Intelligence (AI) instrument was “a spacecraft surrounded by galaxies.” The result, as pictured below, was an image of a vessel suspended in space that appears to reflect the universe around it – largely true to the sign.
For MidJourney founder David Holz, a powerful aspect of generative AI is its “ability to integrate with language”, where we can “use language as a tool to build things”. In simple words, generative AI uses commands from the user to generate novel images based on datasets learned from different sources over time.
The rise of the text-to-image generation has also raised philosophical questions over the definition of ‘artist’.
British mathematician Marcus du Sautoy argues in his book The Creativity Code (Art and Innovation in the Age of AI), 2019, “Art is ultimately the expression of human free will and unless computers have their own version, Till art is art made by one. Computers can always be traced back to the human desire to create.” He says that if we build a “brain” into a machine, perhaps it will offer a glimpse into its ideas. “But we are still conscious of creating code.” are a long way off,” concluded Du Sautoy.
Likewise, notes Holz, “It is important that we do not refer to this as an AI ‘artist.’ We think of it more as using AI to enhance our imagination. Be about but about imagination. We are asking, ‘what if’. AI enhances our imagination.”
Midjourney allows its users to feed in their prompts to its Discord servers and then generates four images that resemble text. The user can choose to explore more variations and pick the perfect fit for a higher quality image. The bot entered open beta last month, giving users a certain number of free trials to bring their imaginations to life. The generated images can also be molded into NFTs, for which, until recently, MidJourney charged royalties.
“It’s a huge community of about a million people all drawing pictures together, dreaming and being separated from each other. All signs are public and everyone can see each other’s images… It is very unique,” Holz told indianexpress.com.
Holz co-founded Leap Motion in 2010, a hand-tracking motion capture user-interface company, and was featured in the Forbes 30 Under 30’s 2014 list. He now runs a small self-funded research and design lab, MidJourney, which, along with 10 other collaborators, is exploring a slew of diverse projects, including AI visualization tools.
Describing the response the AI bot has received, Holz says, “Many people are very happy and find using the product a deeply emotional experience. People use it for everything from projects to art therapy. There are also people who always have some things in their mind but they were not able to express it before. Some people have conditions like aphasia, where the brain cannot imagine things, and they are now for the first time in their life. Using bots to visualize. There are so many beautiful things happening.”
The bot also takes care to prevent misuse of the platform to create objectionable images. The Community Guidelines urge users to avoid using signs that are “inherently abusive, offensive, or otherwise abusive” as well as generating “adult content or gore.” MidJourney also uses moderators who monitor and warn or ban people who violate policies. It also has automatic content moderation where certain words are restricted on the server. AI, too, learns from user data, Holz explains. “If people don’t like something, it produces less than that.”
I pored over the MidJourney bot during a cursory glance through my Twitter feed, where I saw user psychedelic’s somewhat rendition of post-apocalyptic Delhi.
Having previously worked with AI bots like Disco Diffusion and Crayon, an interesting aspect of MidJourney’s exploration was looking at how different AIs would react to the same text. The images below show the results generated with the same indication, ‘Cities during monsoon rains’, by MidJourney, Disco Diffusion, a free-to-use AI tool hosted by Google Colab, and Crayon, previously known as DALL -E was known as Mini.
While Crayon throws up relatively realistic images, Disco Diffusion shows surreal, impressive results, and MidJourney sits somewhat in between the two.
According to Holz, the middle voyage can be understood as a “playful, imaginative sandbox”. “The goal is to give everyone access to that sandbox, so that everyone understands what’s possible and where we are as a civilization. What can we do? What does this mean for the future?”
Holz dismissed fears that AI is here to “replace” humans or their jobs. “When computer graphics was invented, there were similar questions – will it replace artists? And it doesn’t. If anything, computer graphics makes artists more powerful,” he says.
Holz says, “Whenever we see something new, it tries to figure out if it is dangerous and we treat it like a tiger. The AI is not a tiger. It is actually like a big river of water. A tiger is dangerous in a very different way than water. Water is something you can build a boat for, you can learn to swim, or you can build an electric dam. It tries to eat us It’s not angry at us. There’s no emotion or feeling or thought in it. It’s like a mighty force. It’s an opportunity.”