Proposed RFF AI Policy

splitimageview

splitimageview

Administrator
Local time
10:38 AM
Joined
Aug 30, 2007
Messages
10,389
Location
Texas Hill Country
View My Gallery
The RFF staff is currently discussing the following policy wording.
  • No AI-generated or AI-rewritten content.
  • No AI-created images or images modified with AI elements.
  • In other words, do not post AI-generated or AI-assisted content.
  • Basic spelling, punctuation, or translation assistance that does not alter a member’s original wording is not the target of this rule.

The goal is to preserve the forum as a place for genuine human interaction and camaraderie — otherwise it’s like meeting at the pub 🍻and having everyone reply by quietly handing each other AI-written notes. 😆

Everyone’s input is appreciated. Please post comments or criticisms.
 
splitimageview said:
images modified with AI elements.
Click to expand...
Please clarify this concerning post processing using AI tools that do normal PP tasks like auto exposure, contrast, grain, cloning, dust removal, textures, etc. I use Luminar Neo and it has quite a few tools labeled “AI” but most of them just simplify what I could do if I was smarter or had a lot more time to tinker with my photos. It has a tool to add more interesting skies, for instance, but that doesn’t seem right to me and I don’t use it. That may be what you are getting at here, but how do you police that? The quote reads as if only SOOC is preferred. Even some newer cameras use tools to mimic styles that probably could be labeled AI.
 
I think an important distinction has to be made between "AI" and "AI".

This policy seems geared towards GenAI - whether in text form (ChatGPT, Claude, etc.) or image/video (Midjourney, Sora, etc.). This is what most people refer to as AI - and what a lot of people hate with a passion. As soon as I see anyone say "I asked ChatGPT and it said..." I just skip over it. If you haven't done the thinking, I won't do the reading. Simple as that.

But since the rise of GenAI and the related bubble, every piece of automation or machine learning is now branded as "AI" to ride the hype train. That'd be your auto-exposure, dust removal, and cloning tools. I assume that once the bubble pops, every company will quietly rebrand these as "machine learning" again to get rid of the association with GenAI.
 
Good question!

Tools for denoise, sharpening, etc may be labeled as "AI" by some software as @Coldkennels states (and @mrtoml as well.)

"AI-generated" simply means that an image (or an element in an image) that never previously existed so otherwise not possible to be photographed; instead, the subject or element was generated by an AI prompt.

A distinction should be made between editing a photograph of a real subject (or element) and generating a subject (or element) using AI. Adjustments to exposure, color, contrast, dust removal, or other edits applied to a photograph of an actual scene remain photography.

AI generation, by contrast, creates content that was never photographed in the first place. The policy is intended to address the latter.
 
Thank you - that is encouraging.

A further suggestion for your consideration:
- Prevent machines from training on RFF content, if that is possible/affordable.

On another (but related) topic, if that is not already the case, please mandate the use of passkeys/2nd-factor authentication.
 
The draft proposal sounds like a good starting point: No doubt there will be ongoing clarifications needed for situations including synthetic bokeh generated by mobile devices and the like.
 
A lot of today's cameras rely on subject recognition,made possible by AI based machine learning.
You could interpret this as: "AI-assisted content"
For macro I use the Olympus Hand Held Focus Stacking HHFS methode, also Ai based.
What is your position on those advanced technologies? (with more to come in the future?)
 
JeffS7444 said:
The draft proposal sounds like a good starting point: No doubt there will be ongoing clarifications needed for situations including synthetic bokeh generated by mobile devices and the like.
Click to expand...

Portrait-mode blur may use AI machine learning to estimate depth, but it still requires a photograph of a real scene. That’s fundamentally different from generative AI, which creates subjects and content that were never photographed.
 
NLSparks said:
A lot of today's cameras rely on subject recognition,made possible by AI based machine learning.
You could interpret this as: "AI-assisted content"
For macro I use the Olympus Hand Held Focus Stacking HHFS methode, also Ai based.
What is your position on those advanced technologies? (with more to come in the future?)
Click to expand...

Same as above, is the image generated from a prompt?
 
I think it's a clear and sensible rule. I would be happy for a ban on copy-pasting a wall of AI-generated text. I've seen way too many posts and threads (elsewhere) which start out with something like "I thought about this thing and asked ChatGPT about it, and it said this."

I will never read that. There's almost literally nothing that will get me to click away faster than having AI "research" and "summarize" a topic that human beings are talking about.
 
agentlossing said:
I think it's a clear and sensible rule. I would be happy for a ban on copy-pasting a wall of AI-generated text. I've seen way too many posts and threads (elsewhere) which start out with something like "I thought about this thing and asked ChatGPT about it, and it said this."

I will never read that. There's almost literally nothing that will get me to click away faster than having AI "research" and "summarize" a topic that human beings are talking about.
Click to expand...
100% agree, and my solution to date is simply to add any poster that does that to my ignore list, which is slowly getting longer...
 
splitimageview said:
The RFF staff is currently discussing the following policy wording.
  • No AI-generated or AI-rewritten content.
  • No AI-created images or images modified with AI elements.
  • In other words, do not post AI-generated or AI-assisted content.
  • Basic spelling, punctuation, or translation assistance that does not alter a member’s original wording is not the target of this rule.

The goal is to preserve the forum as a place for genuine human interaction and camaraderie — otherwise it’s like meeting at the pub 🍻and having everyone reply by quietly handing each other AI-written notes. 😆

Everyone’s input is appreciated. Please post comments or criticisms.
Click to expand...
Thank for this clear statement, of which I wholeheartedly approve.
I have no idea why so many people are responding with worries about editing, that really is not the point and I thought you made that very clear.
This may not be easy to moderate, but if you can enforce this policy then once again this will be my favourite forum.
Thank you for the effort that has gone into this.
 
nightlight said:
Thank you - that is encouraging.

A further suggestion for your consideration:
- Prevent machines from training on RFF content, if that is possible/affordable.
Click to expand...

That horse has truly left the stable. As RFF is public facing, it will have been sucked dry by AI already. Anything implemented after this date will protect the site, but before this? Nah.

nightlight said:
On another (but related) topic, if that is not already the case, please mandate the use of passkeys/2nd-factor authentication.
Click to expand...

I agree with this, too. 2FA is the way.
 
Archiver said:
That horse has truly left the stable. As RFF is public facing, it will have been sucked dry by AI already. Anything implemented after this date will protect the site, but before this? Nah.



I agree with this, too. 2FA is the way.
Click to expand...
2FA came up before and I and many others were clear we would not go with that under any circumstances. This is a forum, not a bank account.
Please can we stick to the subject. It is important.
 
trix4ever said:
I have no idea why so many people are responding with worries about editing, that really is not the point and I thought you made that very clear.
Click to expand...
It was the phrase 'AI-assisted content' that concerned me with regard to images. That could include masking, blurring, focus stacking etc. But I think everything is clear now.
 
I genuinely think this is a step towards the right direction. With Ai having creeped in every form of social media, we need also spaces free from it.

Ai is the reason I abandoned fb and Instagram - just too much rubbish being generated very quickly and easily that I really need to seive through a lot of it to come across something I find interesting.
 
To quickly address 2FA even though it is not part of this topic:

2FA is an option available to members. It does serve to help eliminate the problem hijacked accounts, which occurred several times last year, resulting in classifieds being posted from those hijacked accounts and members getting scammed. The scammer would take over a long-term RFF account with many posts, lending apparent legitimacy to the classified.

We have implemented several changes to help prevent this, 2FA would add one more layer of security.

Now back to AI discussion!
 
mrtoml said:
It was the phrase 'AI-assisted content' that concerned me with regard to images. That could include masking, blurring, focus stacking etc. But I think everything is clear now.
Click to expand...

Revised for clarity:
  • No AI-generated or AI-rewritten content.
  • No AI-created images or images modified with AI-generated elements.
  • Basic spelling, punctuation, or translation assistance that does not alter a member’s original wording is not the target of this rule.
 
You must log in or register to reply here.
Back
Top Bottom