New AI-Powered App Lensa is Causing a Stir Within the Art Community
AI and Data Science Newsposted by ODSC Team December 6, 2022 ODSC Team
Lensa, the AI-powered app that promises users the ability to morph photos into stylized art has hit the internet by storm. Though not a new app, its latest “Magic Avatars” feature has gone viral, with users on Instagram, TikTok, and Twitter, sharing the hashtag #Lensa to show off their creations. But not everyone is excited at the prospect of Lensa, and this is compounded by that app’s surge in popularity. Last month, Lensa saw over 1.6 million downloads, an increase of 631% from October. This has opened the door wide open about the pros and cons of art created by AI.
For artists, they fear being put out of work by programs like Lensa that can create art much faster. The issue is training data, as AI programs train more and more with existing photos and pieces of art, their ability to create art on its own increases. Meanwhile, an artist would need hours, days, or more, depending on the project. An AI program such as DALL-E could do the same or similar work in a fraction of the time. For many artists, especially those who aren’t well known, this could become a barrier to entry within the field.
Though many might feel that a program cannot fully replace the human component, it might not have to. With the work in art, video, games, and other art-heavy industries, it could act as a massive labor disruptor. Displacing artists who the AI programs depend on for their original and ongoing training datasets. It seems that this concern hasn’t completely fallen on deaf ears. For example, in DALL-E’s case, it learned from stock images provided by Shutterstock. In turn, Shutterstock paid artists whose work was copied by the program in an effort to avoid ethical concerns related to the use of the photos.
But, not every company is taking Shutterstock’s lead and attempting to compensate artists for their original work. Due to copyright law lagging behind technology, The U.S. Copyright Office has taken the position that AI art created without the element of “Human authorship” doesn’t qualify for protection, and thus cannot be registered. Though this is being challenged in federal court, there is nothing in legal guidance when it comes to the use of AI to produce art specifically designed to mimic the styles of living or dead humans.
Below is a Twitter thread on the subject, where a number of artists, without their consent, had their work used by AI for training. In essence, copying their art style:
Before I go to bed, I wanted to say this. I love the art community deeply. When AI artists steal/co-opt art from us I don’t just see art, I see people, mentors and friends. I don’t expect you to understand. #artcommunity #aiart pic.twitter.com/5z77aqI3aK
— Jon Lam (@JonLamArt) December 5, 2022
With all of these ethical concerns surrounding AI and art, it’s clear for now it’s here to stay. How this will affect the world of art is still unknown. But their concerns are mirrored in other sectors, such as content creation, copywriters, etc., who are witnessing the birth of programs such as Anyword & Jasper which are learning for writers. Though AI has been promised as a means to enhance existing human ability, will it? Or can it become a long-lasting labor disruptor – upending generations of jobs, similar to the industrial revolution and how it affected the agricultural community?