Why am I doing this? Why would I pay $8 for 50 different abstract, impressionistic, puzzling portraits of myself? Am I that narcissistic? Are we all?
Our Instagram and other social media are currently flooded with these high-resolution interpretations of our faces and various images Fool, also known as Magic Avatar. They all come from the Lensa app, an AI image generation engine that can put some wild, often fantasy-driven interpretations of any image we feed it.
There are other similar platforms like Reface, which lets you put your own face into famous movie clips. Everyone has been doing this for a while. I turned myself into Tom Cruise (opens in a new tab).
The app was free (with lots of ads, if I recall), and eventually we all decided it might not be a good idea to upload our images to some random app developer.
Lensa fills that space, but its product pretty much requires you to pay for it, first through a hard sell through expensive subscription fees and then more casually with offers to pay in batches for AI images.
It astounds me that there are tons of people paying to have their images digested and then spit out as compelling AI artwork. FOMO is so powerful here that everyone (even the now-embarrassed reporter) caves in. My adult son paid close to $15 for a batch of pictures of his girlfriend.
I’m not proud
Why am I doing this? Two reasons. I’m tired of seeing other people’s Lensa Magic Avatars on my feed, and I really need something fresh for Instagram too. I’m kind of that way with my social feeds, always trying to keep the pipes full, for reasons I’m having a hard time explaining here.
It wasn’t too hard to add my AI images to Instagram’s growing collection of Lensa-generated portraits. The app is available for free on iOS and Android. Opening it gets you a $49.99 subscription deal that you can “cancel anytime.” I personally hate apps like this, the ones that tease you with a really cool feature but require a hefty amount of cash up front (I might as well be cheap).
Like many other applications, Subquote is a kind of foreground. If you ignore it, the app immediately forwards you to the pay-per-view section, where you can buy between 50 ($7.99) and 100 ($14.99) of AI-generated Magic Avatars without subscribing to anything.
Now, I can’t remember the last time I paid $8 for the privilege of using a single app feature, but that’s the siren call for those beautiful portraits. I rationalized the cost in my head, “Come on, you’re going to pay that much for two Krispy Kreme donuts and a glass of water,” which is something I did buy last week.
Afterwards, the app asks you to upload 10 to 20 photos of yourself in various poses and expressions. I decided to peruse my iPhone’s “Selfies” folder, where I found a bunch of suitable photos. I uploaded a batch, but Lensa rejected some as unusable (it didn’t say why), so I picked some more. In hindsight, I should have chosen far fewer side shots and grimacing images. As with most AI imagery systems, you get what you put into Lensa.
Overall, I’m happy with the results (you can see a few examples above) but I left uneasy and feeling a little bit gamed. Before uploading my photos, there are a few questions I should have asked, many in fact.
The day after Lensa provided me with 50 4K AI images (they stay in the app, but you can download your favorites in standard or 4K resolution), I sent Lensa developer Prisma Labs A list of questions:
- Do you store photos that are uploaded to you? If so, are they encrypted?
- When will you launch the Magic Avatar section?
- How exactly does AI generate images?
- How many people uploaded their photo?
- Why would you prepay for a subscription when people can buy 50 images for only $7.99?
- How do you address people’s concerns about uploading images without permission?
- Have you heard of some being used to generate adult content?
In some ways, you can apply any or all of these problems to a virtual AI image generation system. As exciting as all these AI systems are, they really feel like black boxes. For some, we simply entered text prompts, but we still don’t understand how they generate art.
For example, the heavy lifting behind Lensa’s Magic Avatars was done by Stable Diffusion, a free and open source image generation platform – one of many that have been developed Accused of appropriating artist’s work (opens in a new tab) Train its artificial intelligence model.
But that’s just one of my concerns. The other is what happens to the imagery we are so willing to feed it? For example, how do we know that the images we upload are not being used to further train the AI? I don’t see any way to rule out this possibility.
Nor are there any prompts warning you that you shouldn’t upload someone’s image without permission.I’ve heard people using this system Generating pornographic images of unsuspecting people (opens in a new tab) without their consent.
Where does our original imagery go when Lensa’s work is done? No promises that Lensa is deleting them, I have to assume they keep them all on file. We can only hope they are at least encrypted.
Prisma Labs Response
I do have some good news.Although Prisma Labs didn’t include a lot of information in their app that I thought was necessary, nor did it directly answer my question, it did point me to a Lots of FAQs (opens in a new tab) This does solve some of those problems.
Regarding the question of what Lensa does to our images, Prisma Labs wrote that once processing is complete, they are deleted from Prisma Labs’ servers. However, the FAQ also contains this language:
“We store avatars for as long as we serve our users. Please note that a new feature is currently in the works that will allow users to permanently delete their avatars from our servers.”
In other words, your original photos are gone, but what about those images created by Lensa? Prisma Labs still owns them. By the way, there is no mention of data encryption.
A note on image ownership not covered in the FAQ: The images you upload should be your own or one you have access to. It does address “sexualized” imagery, but mostly the AI tends to sexualize female characters. What’s more, Prisma Labs added: “Our terms of use absolutely prohibit the intentional use of the app to create explicit content and require Lensa users to comply with that requirement.”
It’s unclear how Prisma Labs is policing that particular rule, adding: “Unfortunately, all of these efforts have yet to make AI absolutely safe from biased content and explicit imagery. Therefore, we stipulate that the product does not It is suitable for use by minors and warns users of potential content risks.”
I can’t reasonably tell you not to try this app, especially after I go ahead and do it. Before you do, however, make sure you’re comfortable with Prisma Lab’s answers (and lack thereof) to any of these questions.