A Journey into Midjourney
The conversation around artificial intelligence continues to grow stronger than ever, and AI-powered chatbots are becoming more and more common in the public sphere, questions about the limits and the potential applications of Artificial Intelligence are popping up everywhere. OpenAI, Microsoft, and even Snapchat have rolled out AI chatbots that have made waves across tech industries. While products like GPT-4 promise the most advanced AI assistant experience to date, these chats tend to have crucial limits: they’re text based.
ChatGPT and Bing are chats that receive text and return text as an output. One thing these chats cannot return directly is an image. There are some ways around this, but for the most part, AI image generation falls in the court of a few other programs like DALL-E 2, or Midjourney.
AI Generated Images
In short, Midjourney is an AI software that turns user-entered text into images. Developed by the research lab of the same name, Midjourney has been trained on a vast collection of images and data, as well as on language interpretation, in order to turn simple or complex prompts into beautiful, original images. As the development progresses, Midjourney is able to interpret more complicated and layered prompts, and is able to capture smaller details in its image generation, allowing for a more accurate and precise depiction of the user’s input.
Importantly, while there are other AI image generators available to the public, Midjourney puts a focus on artistry. Other powerful AI generators like DALL-E also do a fantastic job at generating accurate representations of the input, but often focus on direct accuracy. In other words, without the “proper” stylistic language in the prompt, these images might come out “boring” or “dull.” Midjourney defaults and biases towards highly stylistic images. According to its founder, this is just because Midjourney is meant to be easy to use. Anybody should be able to create beautiful images at the click of a button.
A Bit of Controversy
AI image generation has been an impressive leap in technology, to be sure. Whenever leaps like this happen, there is some pushback to expect. With the rest of the AI news, pushback has looked more like a fear of displacing workers through automation, or potentially enabling harm by lowering the bar of entry for malicious acts via AI assistance. With image generation, this displacement fear is around too, but also a more nuanced fear about originality.
As far as the fear that AI will push artists, animators, filmmakers, and more out of their jobs, while the fear is understandable, a look into the past should bring us a bit of comfort. For one, while much of the AI image generation is very impressive, there is still quite a ways to go before the technology can functionally “replace” human artists, especially as far as animation and filmmaking go. More than that, however, the introduction of new tools into the art world has bred controversy before, and artists are still here dedicated to their discipline and passion. When animation turned digital, many artists picked up a digital pencil and continued to do what they loved doing. Learning to incorporate the AI tools into now-traditional art styles and methods is, at least conceptually, no different.
A more nuanced concern is just how original these generated art pieces really are. On one hand, the algorithm is quite literally trained on other people’s art styles and techniques. Meaning that although the generated pieces are technically original, they are explicitly pulling from the style of other real artists—and usually uncredited. On the other hand, this is also just how real, human artists learn how to make art. We learn from others and from our communities. Our personal art styles become a melting pot of the art we enjoy and experience. Whether this constitutes plagiarism is still an open discussion in the community.
A clearer example of plagiarism is possible, of course. AI generated images do lower the barrier of entry for image creation. While an artist can practice and train to recreate another artist’s style and technique down to the brushstroke, now virtually anybody can do so with a few keystrokes. If one were to do so explicitly, without the consent of the original artist, and then pass the generated copy off as an original for a profit, there’s a clear ethical problem to consider.
There are more ethical problems like this one to consider, and they all deserve serious consideration and contemplation. Whatever conclusion we reach, however, it is important to remember that, these concerns are concerns about bad actors. Lowering the barrier of entry for malicious acts is not a trivial concern, but we should also grapple with whether to concentrate on the tools that bad actors might use, or the bad actors themselves.
Applications and Uses
To be clear, the excitement around Midjourney is not all controversy. There are also some very real and exciting applications to explore that go beyond just generating beautiful images for the fun of it. One of the biggest applications you’ll see floating around is the generation of images and content for marketing or advertising purposes. The careful details and text of marketing images probably still need to be added and adjusted by a human hand and eye. Especially when the pictures used in banners, posters, and other mediums are secondary and likely to be pulled from some royalty-free collection. Midjourney serves as a quick and easy way to generate attractive, relevant, customized, and original images to include in the document.
Another useful application is the generation of proofs-of-concept. For animations and video games, for example, the AI tech isn’t at the point where it can replace animators or even be a stand-in for the illustrative process. It can certainly be a useful way for a writer or designer to give the illustrator an idea of what they want. It can be a way to search for inspiration when an artist is stuck taking a few key details into the final concept.
There are many exciting and useful applications like these to explore with Midjourney and other AI image generation services. As with any technology like this, concerns about malicious use are present and perfectly valid, and it is important that we hold each other accountable if we want these tools to be mediums for creativity and expression rather than harm. One thing is for sure: if generating stunning, complex pieces of art depicting just about anything you can imagine in a matter of moments and at the click of a button seems like the future to you, you’ll be excited to know that the future is now.
Living Pono is dedicated to communicating business management concepts with Hawaiian values. Founded by Kevin May, an established and successful leader and mentor, Living Pono is your destination to learn about how to live your life righteously and how that can have positive effects in your career. If you have any questions, please leave a comment below or contact us here. Also, join our mailing list below, so you can be alerted when a new article is released.
Finally, consider following the Living Pono Podcast to listen to episodes about living righteously, business management concepts, and interviews with business leaders.