In recent weeks, the lensa photo and video editing app has shot into social media prominence, after adding a feature that lets you create stunning digital portraits of yourself in contemporary art styles. It does that for a small fee and the effort of uploading many different pictures of yourself. Text-to-media artificial intelligence technology left the labs and started colonizing our visual culture, and Lensa may be the most slick commercial application of that technology to date. Lensa is basically prompt engineering on top of Stable Diffusion packaged as an app that's rumored to be making $1M/day from in app purchases. This is just one example of a new generation of apps that will turn prompt engineering into productivity and entertainment tools. pic.twitter.com/LiVxInklD5
There is a different kind of fire among the art community as a result of it. Kim Leutwyler said she recognized the styles of particular artists in the portraits.
The ease with which individual artists' styles can be mimicked has sounded warning bells since Midjourney, OpenAI's Dall-E, and the Comp Vis group's Stable Diffusion burst onto the scene.
The intellectual property of artists has been compromised. Has it been done?
It's not as far as the law sees it.
It is possible for non- computer-scientists to understand text-to-media artificial intelligence.
I'm cropping these for privacy reasons/because I'm not trying to call out any one individual. These are all Lensa portraits where the mangled remains of an artist's signature is still visible. That's the remains of the signature of one of the multiple artists it stole from.
A 🧵 https://t.co/0lS4WHmQfW pic.twitter.com/7GfDXZ22s1
— Lauryn Ipsum (@LaurynIpsum) December 6, 2022
It's important to understand how artists' individual styles can find their way into and out of the black boxes that power systems like Lensa.
The front-end for the Stable Diffusion model is called Lensa. The system used to power it's creative output is called "Latent Diffusion".
There is a key word here. Latent variable is a quality that can't be measured directly but can be inferred from things that can be measured
When Stable Diffusion was being built, machine-learning algorithms were fed a lot of image-text pairs and taught themselves a lot of different ways to connect them.
None of the knowledge base is directly intelligible to humans. Stable Diffusion sees a universe of numbers and connections even though we might see "modernism" in its outputs.
The numbers generated from the original image-text pairs are the basis of all of this.
The system can plot a course through the enormous sea of possible outputs by typing in meaningful prompt.
As an example, take the picture below. The terms "digital art" and "art station" were included in the text prompt.
Stable Diffusion was trained to associate certain words with certain qualities. An image that would fit on ArtStation was created.
If Stable Diffusion is a text-to-image system where we navigate through different possibilities, then Lensa is quite different. One of the biggest innovations of the company is streamlining the process of text inversion.
Stable Diffusion's existing knowledge base is injected with user-supplied photos to teach the system how to capture the user's features. The process is far from a streamlined one.
Although you can't push the images on Lensa in any particular direction, the trade-off is a wide variety of options. The images do not contain any actual work from the other artists.
While individual artworks are subject to copyright, the style and ideas behind them are not. The Bortin case in the US established that the law does not apply to art styles.
The fact that art styles and techniques can be used in this way is upsetting for artists. Legislation may be put under pressure to adapt to the new technologies as artists feel ripped off.
The future looks difficult for artists who work on small scale jobs.
Even though it is easy to make an artwork that looks good using artificial intelligence, it is still difficult to create a specific work with a specific topic and context. The personality of the artist is an important context for their work regardless of how Lensa changes the way art is made.
It is possible that artists will need to invest more effort in publicizing themselves if they want to be taken seriously.
It's early days, but it's going to be a difficult decade for producers and consumers. One thing is certain, the genie is gone.
Brendan Paul Murphy is a lecturer in digital media at the university.
Under a Creative Commons license, this article is re-posted. The original article is worth a read.
All Rights Reserved © 2024
