Well, that's what I thought/was misled to believe, but it's not so. Technically, you can 3D print photographs, but the result is probably going to look like shit (unless you're trying to print an image in some sort of mounting, like a frame). The crux of the issue is that a photograph is a 2D image and a 3D print is, well, 3D. There's a missing z-axis that needs to be created.

Turning a 2D image into a 3D model that can be printed is not a simple task. Sure, there are AI tools (like the DPT Depth Estimation tool) that can turn a single photo into a 3D model, by creating a light map based on luminosity values as a proxy for distance from a focal point. However, for best results, your subject needs to be backlit and against a white background, in a high-quality image. Even then, you'll tend to get a model that's pretty foreshortened/shallow, hollow and contains a number of artifacts you'll not want included in the printed result. (I'd like to show you examples, but the  GLTF files are too large to upload and I don't think Pub0x even supports them anyway. The best I can do is some screenshots to give you an idea.) That means you'll need to edit the model in Blender (or some other modeling software) to get rid of the artifacts and add supports so that what you're printing doesn't fall over before the process is complete. (Gravity is a heartless bitch and certainly isn't kind to flimsy, top-heavy objects, like poor reconstructions of buxom women.)

The front view looks passable, but everything else is a mess. You probably can't even tell which plane is which.

Here's a slightly better model, which has a little more depth (and gives you an idea of from which direction it's being viewed), but it still leaves a lot to be desired:

For best results, you'll need at least three photographs of the object from different angles/perspectives (although six to ten will be better). Unfortunately, I couldn't find any production-ready tools that take this many. (They're all still alpha/beta builds that aren't yet ready for release or general use.) When I do find something that I can use, the most likely thing I'll be able to create/replicate from photos is my old bitch sleeping in her or my bed (and I'm not going to have photos of her from below). Of course, an actual 3D scan (using something like an XBOX Kinect and a potter's wheel or backless stool that can be rotated) would be the best, but it will still give you a model that will need to be refined.

In short, you're better off learning Blender (and/or Godot/Unity 3D) or some other modeling and rendering tool and creating a model from looking at the photograph(s) yourself. In my opinion, this isn't something that AI does satisfactorily quite yet. Maybe it will be a thing that AI does, some day soon. For now, it's not there yet and this is just hype, a promise on which it doesn't deliver. Add Blender to my list of software/technologies I'd like to learn.

In short, I will not be printing 3D models of my AI-generated artwork any time soon, no matter how much I want to have custom doll merchandise to ship to anyone whom subscribes to the top tier of my Patreon. Images and patience will have to do for now, since my first priority is earning crypto and taking fiat from other people (including getting AI image generation functionality added to my Website, now that I've figured out how to use the Hugging Face and Stable Diffusion APIs).