Photography and AI

[Citysnaps]

Two days ago I posted a photograph of a San Francisco beach scene on a Facebook photography group.

My photo received some reactions and comments, which was nice. But there were also a couple comments claiming it was AI-generated.

I’m not losing any sleep over that, though it did get me thinking about photography going forward in a world where AI is getting loads of attention. From the aspect that any photo that looks a bit different or unusual being challenged as being real. As well as the potential proliferation of AI-generated photographs being published, with some being claimed as real.

While that doesn’t affect me personally, it does make me wonder about how photographs and their credibility will be viewed in the future.

Thoughts?

 

[Herdfan]

It is going to be a huge mess. For everyone.

I saw a video of AOC saying some random stuff. Things I know she would have never said. But it looked/sounded about as real as if she were sitting across the desk from me.

In 5 years I am not sure we will be able to know what is real vs fake.

 

[Citysnaps]

It is going to be a huge mess. For everyone.

I saw a video of AOC saying some random stuff. Things I know she would have never said. But it looked/sounded about as real as if she were sitting across the desk from me.

In 5 years I am not sure we will be able to know what is real vs fake.

That's pretty scary. Not necessarily for her in that situation, but in general.

I haven't thought through the consequences and ramifications for different situations, yet. But I feel confident there will be some that are potentially adverse.

 

[ArgoDuck]

A different photographic subject, but this recent research article published in Psychological Science (Miller at al., 2023) relates

Miller’s team tested participants’ ability to correctly identify which were real versus AI-generated human faces. I’ve only glanced through this paper but its methods and analysis seem good.

Their conclusion. “The present study demonstrates a robust AI hyperrealism effect: Remarkably, White AI faces can convincingly pass as more real than human faces—and people do not realize they are being fooled.” (White faces were used as they form most of the AI training data. Non-white AI faces are less convincing).

 

[mollyc]

Adobe has spearheaded a project that allows images to be sort of encrypted, or credentialed, in camera. These keys stay with the image to verify the veracity that it is in fact a real photo, and I think it somehow also track edits being made. The new Leica M11-P has this as a feature, and I've read that Nikon has worked with Adobe also, although it hasn't been implemented into any of their cameras yet. I imagine this will start to become more prevalent and ultimately standard in a lot of camera bodies.

Leica Launches World’s First Camera with Content Credentials Built-in – Delivering Authenticity at the Point of Capture

Industry-leading camera manufacturer Leica is officially launching the new M11-P camera — the world’s first camera with Content Credentials built-in.

blog.adobe.com

 

[Herdfan]

Adobe has spearheaded a project that allows images to be sort of encrypted, or credentialed, in camera. These keys stay with the image to verify the veracity that it is in fact a real photo

How would that work if it is a real photo of a fake photo?

 

[mollyc]

How would that work if it is a real photo of a fake photo?

it verifies the image was taken with a camera

 

[Eric]

Two days ago I posted a photograph of a San Francisco beach scene on a Facebook photography group.

My photo received some reactions and comments, which was nice. But there were also a couple comments claiming it was AI-generated.

I’m not losing any sleep over that, though it did get me thinking about photography going forward in a world where AI is getting loads of attention. From the aspect that any photo that looks a bit different or unusual being challenged as being real. As well as the potential proliferation of AI-generated photographs being published, with some being claimed as real.

While that doesn’t affect me personally, it does make me wonder about how photographs and their credibility will be viewed in the future.

Thoughts?

I get this quite a bit and am typically flattered by it for some reason, maybe thinking the quality is good enough to be considered that way, not sure. I even have drone video footage that gets scrutinized that way as well. However, I think if you have a real body of work people can view it lends credibility to your craft, the lines are definitely getting blurred so I think we'll eventually have to rely on that.

 

[Citysnaps]

I get this quite a bit and am typically flattered by it for some reason, maybe thinking the quality is good enough to be considered that way, not sure. I even have drone video footage that gets scrutinized that way as well. However, I think if you have a real body of work people can view it lends credibility to your craft, the lines are definitely getting blurred so I think we'll eventually have to rely on that.

What I find annoying is the subtext suggesting I'm a fraud and pulling a Santos. OTOH... F 'em.

I did try and engage one commenter in some polite conversation asking why they thought it was AI. The response was, paraphrased: "It doesn't look real. And if it is you should be happy." If? Jeeeez.

How about showing a wee bit of curiosity and first asking a few questions about the photo... how did I make the capture, why does it look the way it does, why does the beach have large static pools of water, etc, etc.

Reflexively coming to the conclusion it's a fake is a cheap shot.

Anyway... I can see more of this occurring in the future now that AI is here.

 

[Eric]

What I find annoying is the subtext suggesting I'm a fraud and pulling a Santos. OTOH... F 'em.

I did try and engage one commenter in some polite conversation asking why they thought it was AI. The response was, paraphrased: "It doesn't look real. And if it is you should be happy." If? Jeeeez.

How about showing a wee bit of curiosity and first asking a few questions about the photo... how did I make the capture, why does it look the way it does, why does the beach have large static pools of water, etc, etc.

Reflexively coming to the conclusion it's a fake is a cheap shot.

Anyway... I can see more of this occurring in the future now that AI is here.

I'm to the point with my posts on both FB and IG that I can't always keep up with the comments but I do try to respond to the ones that count. One thing I've learned is to just let the trolls be trolls and ignore them, if it's something truly hateful I'll block but other than that I just skim by anymore, otherwise I would be arguing with them all day long, just not worth it.