‘Magic avatars’ flooding your social media feed? Thank or blame Lensa AI Filter for it — More about the viral app

Lensa AI has been going viral since it was launched. The cool 'magical avatars' have been flooding social media sites and have taken the internet by storm. However, there are massive ethical and legal questions that surround the ease with which the images of self and others can be morphed to create what techies are calling "soft porn". AI’s Frankenstein's monster seems to be here.

Updated Dec 14, 2022 | 09:05 AM IST

The new AI-Filter app Lensa has taken the social media by storm.

Photo : Times Now Digital
Has your social media feed been flooded with AI-generated avatars of late? Even if not flooded, you must be coming across pictures of your friends which make them look like a vibrant work of art – called " magic avatars ". And for these avatars, you have to credit Lensa AI — the artificial intelligence-powered photo-editing app.

What is Lensa AI?

Created by the United States-based company Prisma Labs, Lensa AI uses artificial intelligence (AI) to create ‘magic avatars’ from your photographs. These avatars could be reimagined versions as anime characters or celestial beings. You will need to download the app — it is available for users on iPhone and Android — and upload your picture. The app will then work its magic and turn the pic into an avatar.

How does Lensa AI do that?

Lensa AI uses a technology called ‘stable diffusion’. Stable Diffusion trains its artificial intelligence to use a massive network of digital art scraped from the internet, mix them and create new ones. Lensa AI stands apart from its predecessor in 2016 by producing more impressive, remarkable and varied versions of the pictures. The 2016 version of the app, which also went viral back then, created painting-like images of the pictures taken on smartphones.

Is it as popular this time too?

Seems like it. According to a report by Sensor Tower, over four million people across the world have downloaded the app and spent nearly $8 million (over Rs 65 crore) creating different avatars, in just the first five days of this month.

So, is Lensa AI a paid app?

Of course it is. While there a few filters which are free, the more advanced ones fall under the “premium” category. The app offers a one-week free trial with limited filters after which the users can opt for a year-long subscription to the app which costs $35.99 (about Rs 3,000).
To generate 50 self-portrait avatars on Lensa AI, you will need to pay $8 (Rs 190). To make 100 ‘magic avatars’ you will need to shell out $12 (Rs 290) and to make 200 portraits, you will need to cough up $16 (Rs 380).

How exactly can you make your ‘magic avatar’?

Once you upload your picture, you can select a category for the type of image and filter you want for you selfie or picture. It would take you about half an hour to see what you would look like as an anime character or as an animated version of yourself against different backgrounds.
What are the concerns around it?
Since the app has gained popularity, it has also been subjected to greater scrutiny about issues ranging from the terms of service and privacy-related concerns to the platform being used to generate non-consensual NSFW images. The other primary concern has been raised by artists is of copyrights. Some of the artists have alleged that Stable Diffusion relies on their artwork to make their own images. However, they are not credited or compensated for their work.

But can AI reallybe tricked?

First, coming to the non-consensual NSFW images part.
TechCrunch, the US-based tech online newspaper, says that it is “way too easy” to use the platform to “generate non-consensual soft porn”.
To prove this, the website used two sets of images. “To verify that Lensa will create the images it perhaps shouldn’t, we created two sets of Lensa avatars: One set, based on 15 photos of a well-known actor; Another set, based on the same 15 photos, but with an additional set of five photos added of the same actor’s face, Photoshopped onto topless models.”
The results were perplexing. While the first set of images was in line with the AI avatars that Lensa had generated in the past, “The second set, however, was a lot spicier than we were expecting. It turns out the AI takes those Photoshopped images as permission to go wild, and it appears it disables an NSFW filter. Out of the 100-image set, 11 were topless photos of higher quality (or, at least with higher stylistic consistency) than the poorly done edited topless photos the AI was given as input,” TechCrunch's Haje Jan Kamps writes.
Not that generating sexualised images of actors and celebrities hasn’t been done in the past. And just because it has been done in the past does not make it right. Everyone deserves their privacy — celebrity or not. No one likes or wants to be made victims of non-consensual sexualized depictions. However, creating realistic images earlier took a lot of skills and expensive photo-editing tools. And even with these specialised skills and tools, it would take some intense hours to create such a picture — most would find the effort not really worth it.
However, the ease with which can create near-photorealistic AI-generated art images by the hundreds without any tools other than a smartphone, an app and a few dollars poses the big turning point and ethical nightmare, writes Kamps.

What does Prisma Labs have to say about that?

According to TechCrunch , when Prisma Labs was asked about these concerns, the company reportedly responded to TechCrunch saying 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.”
When TechCrunch’s findings were submitted to Prisma Lab, the company’s CEO and co-founder Andrey Usoltsev reportedly told TechCrunch in an emailed statement that it is a breach of their terms of use. “The way (TechCrunch’s) experiment was structured points out that such creations can’t be produced accidentally. The images are the result of intentional misconduct on the app, adding that the generation and distribution of such content may incur legal actions as both the US and UK have an act which makes “an act of sharing of explicit content and imagery generated without consent as a crime.”

AI’s Frankenstein's monster

We are not very convinced if it would help those who are or have been victims of their pictures being sexualised without their consent. These victims could be celebrities or everyday people like us. And that thought is very scary.
What makes it scarier is that the AI art generators have reportedly already created (and are creating) scores of pornographic images. These images include unfettered proliferation of other so-called “deepfake” platforms. This creates an unprecedented, unwanted but a significant problem for the law-enforcing authorities. Technology and AI’s Frankenstein's monster of sorts.
While it might be cool to create your avatar that is hyper-real version of yourself, one needs to be very conscious of the kind of image and imagery one would want to portray of themselves and others. If not checked, Lensa AI could bring about the kind of optical nightmare that could lead to a massive ethical, moral and legal quandary.
next story