Samsung’s New AI Image Generation Tool Is A Little Too Good

The pirate ship in Elliott Bay was one thing, but it was a fuzzy little bee that threw me off my feet.

Samsung would like us (and its shareholders) to know that its new phones are the smartest phones ever made, and the Fold 6 I’m reviewing has a new tool called “sketch to image.” Draw a sketch on a photo or a blank page of notes, and it will use generative AI to turn that thing into an image. I shrugged it off as just another AI tool when Samsung announced it on stage at Unpacked — but y’all, it’s really good. So good that it worries me a little.

Using the image-to-sketch tool in a note is pretty harmless: you draw something, highlight it, and choose from a handful of styles like “3D cartoon” and “illustration” to turn your drawing into something more detailed. Your image is sent to the cloud, and after a few moments, you’ll see a handful of options to choose from. The results are usually cute and fun; I’ve taken my two-year-old’s requests and we’ve drawn goofy-looking dump trucks and school buses. Sometimes you’ll get a teddy bear with too many arms, but it’s nothing serious.

It’s when I use Sketch to Image on a photo that things get weird. I’m the world’s worst artist, and this tool has translated my most basic sketches into photorealistic images. The AI-generated elements are convincingly integrated into the photos, scaled and adapted to the environment in a way that makes them hard to spot as fakes.

This is how I got to the bee problem. I took a photo from a dock just south of downtown Seattle with some flowers in the foreground. Since they are close to the camera and my focus was far away, they are slightly blurry. I drew the world’s worst sketch of a bee on one of these flowers, thinking the AI ​​would insert a clear image of a bee, easily giving it away as a fake. Wrong!

The AI ​​bee is blurry, as is the flower it’s landing on. If I didn’t know the AI ​​bee’s origin story, I wouldn’t think twice about walking past this image on Instagram. I’m assuming the photographer either took the shot at the right moment or waited for a bee to enter the frame—things that take skill and patience. They didn’t. In fact, I’m not even sure I can spot the “AI-generated content” watermark in the corner of the image.

I’ve been playing around with the image drawing a lot over the past week, and the results aren’t always so “fuzzy.” Often, they exhibit the telltale signs of a piece of AI generative art: words scribbled in an alien language or strange textures that don’t quite look right. Convincing at first glance, but if you look for more than a second, you’ll notice something is wrong.

Sometimes the content itself betrays reality: I don’t think anyone would believe I saw a huge pirate ship anchored in Elliott Bay or a giant orange cat in a West Seattle intersection. But even when the images are so outrageous that no one could mistake them for real, they seem realistic.

In general, large objects look obviously fake. But it’s very easy to add another car to a photo of a busy road or a sailboat in the distance, and most people won’t know. Aside from this AI watermark, which is easily cropped out, there’s really no way to tell that there’s anything unusual about the image. It’s weird!

I don’t want to exaggerate. Using a sketch on an image is entirely optional, and many people will never even find it in the Gallery app. A blurry bee isn’t going to unravel the fabric of our society. But I think we’re in an increasingly weird place with AI. Sure, you’ve been able to add a blurry bee to an image in Photoshop for ages. But putting that capability into the the same device you use to take the photo and distribute it That’s another thing. The capabilities and accessibility of generative AI tools are outpacing our common understanding of what can be real and what can be fake when you scroll through Instagram.

Personally, I find it most bizarre when I show this feature to my child. He will grow up knowing that with the push of a button you can transform a crude sketch into something more polished. Or, with minimal effort, you can spice up a photo of train tracks by adding a train. Is this a good or bad thing? I have no idea, but I definitely sense a dissonance between how I viewed art as a child and how he will view it.

None of this has stopped me from having fun turning sketches into images. The results of generative AI are pretty hilarious in their honesty, like when I tried to add a green monster poking its head out of the Puget Sound and it interpreted my drawing as a huge green polar bear with rippling muscles standing on the shore. Or when it turned a stick figure sketch into a life-size stick figure, with a shadow on the ground below.

Is the definition of photography changing before our eyes? Is our understanding of truth in images being transformed at an incredibly precarious moment for our democracy? Yes, but also, I took a picture of a rabbit and AI allowed me to put a little top hat on its little head. What a time to be alive.

The sketch image is available on the Galaxy Z Fold 6 and Z Flip 6. Samsung hasn’t publicly stated whether it will make the feature available on other Galaxy phones, but given the company’s track record of aggressively expanding Galaxy AI to previous-gen models, I think it’s extremely likely. Samsung has also committed to bringing AI features to 200 million phones this year alone. If a fuzzy little bee is any indicator, I’d say things are bound to get a little weird when that happens.