In the article "Is Lensa AI Stealing from Human Art?" it discusses a new app called Lensa which is a photo and video editing app which allows people to generate really detailed and aesthetically pleasing portraits of themselves for a small price. The problem is that many popular artists are claiming they recognize their styles within Lensa's various portraits. Though artist's feel stolen from according to existing copyright law nothing is being violated, even though in some portraits the artists signatures are visible on the bottom right. The article dives deeper into the way Lensa runs to figure out what is going on. Apparently Lensa uses a system named Latent Diffusion to fuel it's outputs. The Latent variable of Lensa cannot be directly measured but can be deduced from other variables that can be directly measured. The whole entire process involves some very complex equations that are triggered when the user inputs text prompts such as "Digital Art". Lensa is also different from other AI since it takes inputs from the user in the form of their pictures. Since the artists original images are being changed in Lensa to suit the needs of the user, all that is being stolen is their art style which is not protected under copyright law. The article predicts the future is going to be hard for artists who design drawings for influencers and businesses but notes that it may be rather easy and inexpensive to create art through AI, it is still challenging to get the right subject and context people desire.
Comments