The Lensa photo and video editing app has gained popularity on social media in recent weeks, after adding a feature that lets you create stunning digital photos of yourself in contemporary art styles.
All it takes is a small fee and the effort of uploading 10 to 20 different photos of yourself.
2022 was the year that text-to-media AI technology left the labs and began to colonize our visual culture, and Lensa may be the fastest-growing commercial application of this technology.
Lensa is basically agile engineering on top of Stable Diffusion packaged as an app that is rumored to make $1M/day from in-app purchases.
This is just one example of a new generation of apps that will transform instant engineering into productivity and entertainment tools. pic.twitter.com/LiVxInklD5
– Dar Obasanjo 🐀 (@Carnage4Life) 8th December 2022
This has lit a fire among social media influencers looking to express themselves — and a different kind of fire among the art community. said Australian artist Kim Leuteller The Guardian She recognized the styles of certain artists—including her own—in Lensa’s photographs.
Since Midjourney, OpenAI’s Dall-E, and CompVis Group’s Still Diffusion burst onto the scene earlier this year, the ease with which individual artists’ styles can be imitated has been ringing alarm bells.
Artists feel that their intellectual property—and perhaps a bit of their soul—has been compromised. but have
Well, as long as existing copyright law looks at it.
If this isn’t outright plagiarism, what is?
Text-to-media AI is inherently more complex, but it’s possible for us non-computer scientists to understand the concept.
I’m doing this for privacy reasons/because I’m not trying to impersonate any one person. These are all lens paintings where the curved remnants of the artist’s signature remain. These are the remains of a few artist signatures that were stolen from him.
One 🧵 https://t.co/0lS4WHmQfW pic.twitter.com/7GfDXZ22s1
— Lauryn Ipsum (@LaurynIpsum) December 6, 2022
To really understand the positives and negatives of Lensa, it’s worth taking a few steps back to understand how individual artists’ styles can find their way into black boxes and outdoor power systems like Lensa.
Lensa is basically a streamlined and customized front-end that is freely available for stable diffusion deep learning models. It is so named because it uses a system called latent diffusion to power its creative output.
The word “subtle” is key here. In data science, a latent variable is a quality that cannot be measured directly, but can be inferred from things that can be measured.
When stable diffusion was being developed, machine learning algorithms were fed a large number of image-text pairs, and they taught themselves billions of different ways that these images and captions could be connected.
It created a complex body of knowledge, none of which is directly comprehensible to humans. We may see “modernism” or “car paint” in its results, but stable publishing sees a universe of numbers and connections.
And it’s all derived from complex math that involves numbers created from the original image-text pairs.
Because the system captures both description and image data, it allows us to chart a course through a vast sea of possible outcomes by typing meaningful clues.
Take the following image as an example. The text prompt includes the terms “digital art” and “artstation” – a site that is home to many contemporary digital artists.
During this training, Stable Diffusion learned to associate these words with specific qualities that it recognized in the various arts in which it was trained. The result is an image that would fit well in an art station.
What sets the lens apart?
So if static diffusion is a text-to-image system where we move through different possibilities, the lens looks very different because it takes in images, not words. This is because one of the major innovations of Lensa is to simplify the process of converting text.
Lensa takes user-provided images and feeds them into Stable Diffusion’s existing knowledge base, teaching the system how to “capture” the user’s features so it can stylize them. While this can be done in a regular stable spread, it is far from a regular process.
Although you can’t compress images in any particular desired direction on Lensa, the trade-off is a wide variety of options that are almost always effective. These paintings take ideas from the works of other artists, but do not contain any actual pieces of their work.
The Australian Art Law Center makes it clear that while individual artworks are subject to copyright, the stylistic elements and ideas behind them are not. Similarly, Dave Grossman Designs Inc. in the United States. v Burton established the case that copyright law does not apply to artistic style.
What about artists?
However, the fact that art styles and techniques are now so transferable is very confusing and very frustrating for artists. As technology like Lensa becomes more mainstream and artists feel increasingly left out, there may be legislative pressure to comply.
For artists working on small-scale jobs, such as creating digital exposure for influencers or other web enterprises, the future looks challenging.
However, while it is easy to create artwork using AI, it is still difficult to create a unique work, with a specific theme and context. So unless apps like Lensa change the way art is created, the artist’s personality remains an important context for their work.
It may be that artists themselves need to borrow a page from the influencer’s handbook and invest more effort in self-publishing.
It’s early days, and it’s going to be a turbulent decade for art producers and consumers. But one thing is for sure: the genie is out of the bottle.
Brendan Paul MurphyLecturer in Digital Media, CQUniversity Australia
This article is republished from The Conversation under a Creative Commons license. Read the original article.