AI Chat/GPT Pictures and Usage
-
@Giles This might have been mentioned elsewhere, and I don't wanna be "that guy" but should we not just impose a blanket ban on AI/ChatGPT pics unless it is with a huge "AI" watermark through it. or just ban it altogether.
The problem is people end up copying and pasting and/or sending pics to others and before you know it the original context (already shaky) is lost and people are making decisions based on nonsense or people saying "that looks sh*t" because it's a rather shoddily produced mockup of what something might or might not look like.
Or people just quickly scrolling through threads and see pics and think its sh1t and keep scrolling.IMO product pics should be real, with no exceptions. No doubts, no queries, no caveats.
In the last week or so I've seen fake aging pics, fake snap pics, who knows what else.
No offence to any of the original posters, but its a slippery slope to not knowing what is and isn't real/mockup/meant to be funny etc etc.
I could be miles off it here, in which case no offence intended and carry on lol.
-
@Stringthing75 not a bad thought! Thank you for bringing it up
-
big agree here as well
-
There are quite a few ai image detectors. I'm sure you could integrate that into the forum.
-
The topic has been forked/moved here for further discussion @Giles @Ikeaparty @nixage @sabergirl @Stringthing75
-
Completely agree with what's already been said here, 10 years or so people would say "That's photoshopped!" now it's "that's AI!"
At least there was some level of skill/work when people used photoshop...
-
-
@don_pipone I think this is addressed moreso at fake AI evo pics and such.
IG: electricindigowizard
"Possibly splitting hairs, but I consider @Oaktavia to be the beardy, dank High Magus of this denim game…" @neph93
ISO: IHJ-32 - XL/XXL
-
@Oaktavia I understand but still no one knows where this image might end up. I agree with what @Stringthing75 said, I didn’t think about possible implications when I posted it. Even though it‘s probably not as bad as fake evo it still is fake and should be marked as such.
-
@Oaktavia @don_pipone I wont lie, I've asked ChatGPT to give me fake evo pics on jeans like "what would these look like in 6 months" just to see what it could do... I don't think they look real, and you can normally always tell. But it would be really lame for someone to post them and pretend it's their own evo, etc. It could even be a concern for fading comps
I imagine if @Oaktavia asked to evo his jeans, it would just show them disintegrating and turning to dust
-
@Jord it wouldn’t be lame, it would be hilarious
-
@Jord said in AI Chat/GPT Pictures and Usage:
I imagine if @Oaktavia asked to evo his jeans, it would just show them disintegrating and turning to dust
Now I‘m seriously curious what ChatGPT would make out of it -
as I said, no offence to anyone that posted stuff. remember, all this kind of thing is usually done with good intentions. and to be fair, IH crew themselves might already be finding a use for the amount of times they are asked "hey what would this look like with snaps made of platinum from mars" etc.
I don't want to sound like a luddite, but even if the IH crew were to start knocking out 'official' AI generated pics at the request of users, it still leaves the absolute guarantee of them spreading very very easily far and wide. especially as with AI tools you can easily 'strip' digital watermarks off of pics. It completely dilutes the original intention.
Then you're back to square one. One shirt, and about 25 different images that very quickly people lose sight of what is the actual piece.
especially when some stores use the same photography, or scrape their pics from multiple sources. I can guarantee this will only ever result in a headache for the IH crew. Or to put it another way, there is no situation where it is a good idea to lose control of your source image, within reasonable grounds.So really, the only solution is complete integrity of the image. after all, it's the brands absolute gold standard. you control how your stuff is presented. To a point. We know that, but that's all you can do. If you do less than that then you have to accept the grief that will follow. and no one wants that. there's enough going on to be dealing with a disappointed punter from the depths of Scotland (me) who is wondering why his "grail" didn't come with a chain stitch run off even though it was never meant to. etc etc.
if people want to see what a mockup might look like, then if it isn't physically possible, then feel free to crack on and do it outwith the forum and make your decision based on that.
As has been said before, a lot of time, effort and money goes into producing high quality photography, measurements, service etc and I'd hate to see that go down the toilet when it becomes an AI slop fest. There are a lot of threads and images could be anywhere.
It can be subtle stuff. slightly changing the shade of indigo on a pair of jeans, it then wouldn't be easily noticed by the next person.Also, without stirring any recent emotions, the very fresh 'feedback' around some purple and white stripes and then some pockets being slightly misaligned have shown that if a user - with good intentions - used AI (or any tool) to 'fix' the pockets to show what it 'should' have looked like in their opinion and they send it to someone and then that person orders it and then it doesn't look quite like that then you're just giving yourself a huge headache.
what happens on the wilds of the internet is beyond your control. but this forum is essentially an integral part of the brand.my point basically is if we all adhere - to the best of our knowledge - to keep this completely AI/Photoshop free then it retains the integrity of the item, the photography, the vision and importantly, the expectations of everyone involved. It retains everyone's full confidence. Because you don't need to ask that question. You already know.
No tools required, just a policy that people on here adhere to. no caveats.
re-reading that it sounds like I have unilaterally decided and written the policy but hey, you know what I mean.
Where I work we have very good reason to be in complete control of our "IP" so to speak and if you don't set your stall out super early on how you deal with AI and employees/users use of tools then your own commercial interests can very quickly be exposed to a completely unnecessary and unstoppable level of risk. If you pause for 2mins to think about the proprietary information people routinely type into - for example - ChatGPT on work networks then you'd have zero hesitation in outlawing it lol.
Especially more so for smaller businesses that do not have the resources and can be sunk without trace overnight.I realise I have made quite the leap from a few dodgy pics to the END OF THE WORLD but you get the gist. control what you can control basically. you'll thank yourselves later.
I am not always this boring by the way.
Over and out!
️