Your Facebook and Instagram feeds have undoubtedly been flooded by a barrage of AI-generated profile images in the past few days. The pictures, generated by an app called Lensa, produces a variety of captivating images that the Internet has been loving—but it might not be as harmless as everyone thinks it is.
Launched in 2018, Lensa is a product of Prisma Labs — a company based in Sunnyvale, California, which specialises in AI.
Available on both Apple’s App Store and Google’s Play Store, the app is best known for its wide range of filters for photographs. Many of these filters convert pictures into outputs resembling abstract artworks.
However, the app became a rage only in November 2022 with its AI-based feature known as ‘Magic Avatars.’ The feature allows people to convert their photographs — self-portraits in particular — into what appear as photo-realistic paintings or high-quality anime artworks.
Several users have been using the feature to convert their profile photos into Lensa’s AI-created art. Propelled by its massive use, Lensa AI, as the product is named on the iOS platform, topped the App Store in the ‘Photo & Video’ category on 1 December.
Everything to know about Lensa AI
What is Lensa AI?
Lensa requires a minimum of 10 user-uploaded images for it to process an AI-generated photo-realistic version. The accuracy of the output depends on the number of images provided to the app.
The system uses the open-source network known as Stable Diffusion, developed by a start-up named Stability AI, to create digital art.
Using the app for the first week is free. Users will have to pay USD 7.99 per month for subsequent use. A yearly subscription costs USD 29.99.
‘Magic Avatars’ come in separate packages that can be obtained via in-app purchases, ranging from USD 3.99 to USD 49.99.
How to use Lensa?
Anyone can use Lensa and its artificial intelligence feature on their smartphones.
Once downloaded, users can open the app and click on the ‘Photos’ tab. Once in the Photos section, users can click on a yellow button marked ‘Magic Avatars.’
After accepting terms such as the possibility of inaccuracies, users can click on “continue” and then upload at least 10 of their portrait photos.
As per the app, the photos should ideally be close-ups and not feature children or groups. There are several other recommendations related to the type of photographs that can be fed to the AI.
The app declares that all pictures uploaded “will be immediately deleted from our servers after the Avatars are ready.”
Once pictures are selected, the app asks about the gender of the user. Users will then have to pay for use of ‘Magic Avatars.’ This is where the packages come in. According to reports, users can download around 200 different avatars for USD 7.99 for subscribers at a discount of around 51 percent.
Avatars are ready in around 20 minutes and generated for viewing and saving in multiple themes, ranging from Anime to Cosmic and Pop to Kawaii.
What are the controversies around the app and its artificial intelligence feature?
The surge in the number of users notwithstanding, Lensa has come under fire over the nature of images that artificial intelligence generates.
Writing for Wired, Olivia Snow revealed serious problems with Lensa. Snow said that the AI used her childhood pictures to generate objectionable photographs.
Snow in her report said, “Lensa’s terms of service instruct users to submit only appropriate content containing ‘no nudes’ and ‘no kids, adults only.’ And yet, many users—primarily women—have noticed that even when they upload modest photos, the app not only generates nudes but also ascribes cartoonishly sexualized features…”
As per Haje Jan Kamps of TechCrunch, “Lensa will gladly churn out a number of problematic images” from just a handful of 10-15 photos.
The report was updated with Prisma Lab’s response.
“The company highlights that if you specifically provoke the AI into generating NSFW images, it might, but that it is implementing filters to prevent this from happening accidentally. The jury is still out as to whether this will actually help people who are the victim of this sort of thing without their consent,” TechCrunch’s report stated.
On the other hand, many artists have accused Stable Diffusion — the neural network that Lensa uses — of using their work without permission.
Artist Karla Ortiz told NBC News that instead of “bringing art to the masses” as it claims, what Lensa is “bringing is forgery, art theft [and] copying to the masses.”
In response, Prisma Labs CEO Andrey Usoltsev told NBC News that it was “never part of the company’s mission” to bring art to the masses. Usoltsev also praised the “incredible milestone” that Stable Diffusion set in “democratization of access” to technology.
“What was once available only to techy well-versed users is now out there for absolutely everyone to enjoy. No specific skills are required,” Usoltsev said.
On 6 December, Prisma Labs put out a Twitter thread in response to the concerns of artists.
Seeing plenty of thoughts online about the future of digital art in connection with AI generations, we decided to share some information on how AI generates images and why it will not replace digital artists. 🧵🧵🧵
— Prisma Labs (@PrismaAI) December 6, 2022
“As cinema didn’t kill theatre and accounting software hasn’t eradicated the profession, AI won’t replace artists but can become a great assisting tool,” the company said in a tweet.
“We also believe that the growing accessibility of AI-powered tools would only make man-made art in its creative excellence more valued and appreciated, since any industrialisation brings more value to handcrafted works,” Prisma Labs added.
The company also tried to address privacy concerns around images. Some experts have questioned whether the uploaded pictures get deleted as the company claims.
“It’s impossible to know, without a full audit of the company’s back-end systems, to know how safe or unsafe your pictures may be,” cybersecurity expert Andrew Couts, who is also a senior editor at Wired, told Good Morning America.
In its 6 December Twitter thread, Prisma Labs reiterated that “the user’s photos and the associated model are erased permanently from our servers” as soon as the avatars are generated.
“And the process would start over again for the next request,” the company added.
(Main and Featured images: Prisma Labs/@PrismaAI/Twitter)