If you took a moment to glance through Instagram, Twitter, or TikTok over the long weekend, you probably noticed that a lot of your friends and the celebrities you follow have changed their profile pictures to artistic fantasy portrait selfies in which your cousin or workaholic coworker appears as ethereal fairies, elves, warriors, astronauts, superheroes, and manga-style characters. They didn’t all overnight master the art of illustration or digital creation, but rather, like Taraji P. Henson, Tommy Dorfman, Michaela Jaé Rodriguez, and Jennifer Love Hewitt, they simply used Lensa AI, an iOS and Android app that uses artificial intelligence to process selfies and photos to produce avatars that resemble those of professional illustrators.The outcome is so captivating that, according to early projections from analytics company Sensor Tower, more than four million individuals downloaded the app globally in the first five days of December and spent more than $8 million on Lensa AI during that time.

Its nature and mechanism

The popular image-editing app, created by Prisma Labs, was first released in 2018, but it has gained a lot of popularity recently as a result of significant improvements in artificial intelligence, including the “Magic avatar” feature and various tools like Face Retouch and Magic Correction that “refine facial imperfections” and help create reworked images that are both increasingly similar to and perfect in comparison to the original. The app’s success may be attributed to its simplicity of operation. The first step after downloading the app is to select one of the offered packages, which range in price from $3.99 for 50 photographs to $7.99 for 200 images. After that, you will be required to upload 10 to 20 close-up selfies with a range of backgrounds and expressions. The topic of the generated photos will depend on the gender you choose in the third phase, which might be one of anime, fantasy, fairy princesses, astronauts, rock stars, or superheroes. After a brief wait, you will receive a collection of digital artwork to share with your followers and friends.

Privacy

One of the main concerns of Lensa AI users is that Prisma Labs, which has in the past faced criticism for keeping users’ private information on file, will continue to do so. The business’s CEO and co-founder, Andrey Usoltsev, stated that it is now working on amending its privacy policy “Lensa employs a replica of the Stable Diffusion Model, which trains facial recognition for each unique case using uploaded photographs. This implies that each unique user has their own model “.Usoltsev further notes that users’ photographs are immediately erased from the servers (based in the United States) once avatars are created, although it is best to thoroughly read the privacy policy and terms of use section prior to downloading the program. The app “does not use user-supplied photos… for any purpose other than applying different filters or stylized effects,” “user images are used solely for the purpose of creating their own avatars,” and that data is automatically deleted within 24 hours of the images being processed are all stated in this section. It is unknown if additional private information, including a person’s location, can be disclosed to others. If you would want “access, modification, correction, update, deletion or erasure” of any personal information that Lensa submitted to Prisma, you can send an email to privacy@prisma-ai.com. However, the firm cannot promise that your request will be met.

Why artists have problems with Lensa AI

We could quote Spiderman and argue that “with great power comes tremendous responsibility” because the editing app’s enormous success is accompanied by a number of problems. The first is related to copyright. Lensa AI and Stable Diffusion, the open-source artificial intelligence model it uses for generative images using a sizable LAION (Large-Scale Artificial Intelligence Open Network) and LAION-5B dataset, have been criticized in the past for stealing copyrighted images from millions of artists who publish their work on host websites like DeviantArt, Pinterest, and ArtStruct. These artists have voiced their frustrations and warned users about the sneaky nature of Lensa AI Additionally, the accessibility of these AI-produced works of art could eventually replace the creatives’ work with a less expensive, clickable substitute.

hypersexualization and male gaze

The other main complaint against Lensa is that it creates hypersexualized visuals, particularly when a female user is involved. The terms of use for Lensa are violated when children’s photos or nudity are used to create images. Despite this, many testing have revealed that at least one of the generated content features seductive positions, orgasmic facial expressions akin to those in pornographic manga and Japanese anime, abysmal cleavage, enormous breasts, and even full nudity (which opens up the danger of the creation and dissemination of non-consensual images). The Artificial Intelligence used by the program is based on vast amounts of data, and it uses what it learns to make conclusions and predictions based on data that is frequently biased, sexist, or racist because it is developed by humans. As a result, the results it produces may also be prejudiced. By suggesting princesses to women and astronauts to men, Lensa tends to make sexist generalizations. But more importantly, it frequently suggests a highly sexualized, stereotypical feminine image that is a product of the male gaze as well as a Eurocentric aesthetic. A number of Asian ladies highlighted how AI westernizes their faces or alters their characteristics to match a sexy manga, reducing them to nothing more than stereotypes. Others who identified as black skin pointed out that they almost always it looks lighter.

Regarding Lensa and body positivity

For body positivity, does Lensa represent a step back? Many people are questioning as the app’s popularity rises, but the answer is complicated. The distinguishing characteristics, like with most things, are intention and use, both in terms of expectations and timing. If the app is used as occasional pleasure to experiment with one’s appearance and see oneself as different as one might be with bolder makeup or a new outfit, it still remains pretty innocuous. The trap is the same as the one that underlies Instagram’s influx of glossy, edited photos with impromptu effects. Lensa frequently suggests models with hourglass figures, sculpted bodies that are slender and shapely only in particular areas, enormous, toned breasts, flawless, westernized features, and no flaws. In other words, it presents a wholly unrealistic self, which is OK as long as we are conscious of it and do not put ourselves through stress to make that artwork real enough to go to a cosmetic surgeon to achieve that slender princess or ethereal fairy look, as seems to have happened recently. The risk is internalizing and normalizing that artificial intelligence-reworked portrayal, especially for people who already struggle with physical dysmorphia or low self-esteem.