Midjourney released its external editor on Thursday, « a powerful new tool to unleash your imagination. » The artificial intelligence tool, available to a selected group of users, will allow users to upload their own images, then adjust, edit, and reassemble them in a wide range of artistic styles.
Previously, users could upload a reference image to Midjourney, either through its alpha web application or its Discord server, and then have the generation model use it as a reference to create a new image. However, you could not make any modifications to the source image itself. This changes with the new external editor. With it, you will be able to add specific assets within the image, edit, move, resize, remove, and restore them, as well as completely redesign them in a new style – transforming them from a photographic image to another, for example, from pointillistic to impressionistic to anime. The system is said to also work on innovative logo designs and linear graphics.
The new tool has not been rolled out to everyone yet. « To maintain the same standards of fairness we have put in place across Midjourney, » the company wrote in its announcement blog, making it available only to users who have subscribed to its service for a year or more, or those who have produced 10,000 or more images through the Midjourney platform – essentially, its most loyal customers.
We hope that this policy will prevent a flood of generations that violate copyright, as we saw with xAI’s Grok 2 release. Although one user, Haleem Rasasah, has already moved forward by switching a model in an image using the external editor, highlighting the fact that it is almost impossible to control the images created by artificial intelligence.
You will also, of course, need to meet the site’s age requirements, in addition to « following all applicable laws, our community guidelines, and other policies. » The company also reserves the right to remove content at its discretion and warns that « you may encounter friction with our moderation – innocent-looking claims may be blocked by our filters. »
Midjourney founder David Holz indicated in a message on the company’s Discord server that « all these things are very new, and we want to give the community and human moderation staff enough time to handle them gently… »
Users have already begun sharing their edited works online, with impressive examples as much as they are worrisome. For example, Dreaming Tulpa used it to reshape the Mona Lisa painting in various gothic styles – this being « gothic » as in the Hot Topic gothic mall, not the artistic and architectural innovation period in Europe that lasted from the 14th to the 16th century.
New Midjourney editing feature
Yesterday was the last day where reality could be distinguished from fantasy. I hope you made the most of it! pic.twitter.com/TtPjTn8zjD
— nearcyan (@nearchyan) October 23, 2024
Conversely, another user used it to replace the cake with a set of graphics processing units in the image of Nvidia CEO Jensen Huang, boasting that « yesterday was the last day where reality could be distinguished from fantasy. I (this) hope you made the most of it! We still need to see if these overly resource-heavy accounts will lead to anything more than just distorting classic artworks or generating memes.
Midjourney is certainly not the only artificial intelligence company offering these kinds of services. Gemini, ChatGPT (via Dall-E), Grok, and Copilot can also create and edit images in a similar manner.
