Midjourney will launch a web-based tool that can edit any image using generative AI. The new product will allow, while maintaining the shape of edited objects, to change their textures, colors and other details depending on the user’s text requests. This opens the door not only to greater creative possibilities, but also to possible abuse.
Last week, company CEO David Holtz posted details about the new AI image editor on the Midjourney Discord server. The new tool will allow users to download images from the Internet and use the latest advances in AI technology to edit them. “It preserves the shape of the scene and objects, but redraws all the textures, colors and details according to your request,” Holz explained.
The emergence of such a tool opens up great opportunities for users to create, but also raises serious questions about abuse. For example, the tool could be used to create deepfakes or copyright infringements en masse. While the US still lacks a federal law criminalizing the creation of deepfakes, at least a dozen states have already passed anti-imitation AI laws, and many more are considering following suit.
Holtz is aware of the risks associated with potential misuse of the new tool and said that access to the service will initially be limited to a small number of users. Midjourney develops monitoring systems that incorporate both human oversight and AI to minimize potential violations. However, the specific mechanisms for implementing these restrictions have not yet been determined, and the company is collecting feedback from users to decide who should get early access to the tool.
It should be emphasized that Midjourney has already implemented IPTC Digital Source Type metadata standards to allow AI-generated images to be labeled. However, the company still lags behind other market leaders such as Adobe, which use more advanced image provenance tracking technologies such as the C2PA standard. This technology provides the ability to trace the complete history of an image’s editing, which significantly increases transparency and accountability in the process of its creation and processing.
In August of this year, Midjourney also launched an updated website that allows anyone to create images using AI by registering with a Google account. An initial free trial period is provided, after which users can sign up for a paid subscription to continue using the image generation service on an ongoing basis.
This week, Congress held confirmation hearings for new ministers nominated by new US President Donald…
California-based startup Vast Space has announced the completion of the world's first commercial space station,…
In October, Qualcomm introduced the flagship processor Snapdragon 8 Elite, which received two powerful cores…
Samsung has previously announced the development of functions based on Vision AI artificial intelligence that…
OpenAI CEO Sam Altman announced on Platform X on Friday that work on the large…
US scientists have used an innovative method of combining images from two optical telescopes to…