Glenn2
Well-known
There’s a quote I seem to be using more often, source unknown.
“Science is the signal, politics and religion are the noise “.
“Science is the signal, politics and religion are the noise “.
Ai silliness is taking over the internet
- Thread starter p.giannakis
- Start date
- Latest activity Latest activity:
- Replies 101
- Views 2K
valdas
Veteran
Yes, but unfortunately our lives are probably 90% noise… For some it is 100%.
Dogman
Veteran
‘They Couldn’t Break Me’: A Protester, the White House and a Doctored Photo
President Trump and the White House regularly circulate imagery that has been manipulated by A.I. But the photo of Nekima Levy Armstrong was different.
I would worry less about AI being used for sinister purposes if we were guaranteed there would be some legal penalty for doing so. Doesn't seem to be these days. Anything goes for anyone anytime anywhere.
splitimageview
Administrator
Editing photos or videos to evoke a specific response is nothing new, didn't need AI for that.
For example, when Joe Rogan got COVID, CNN made him look like a green Martian!
For example, when Joe Rogan got COVID, CNN made him look like a green Martian!
Coldkennels
Barnack-toting Brit.
According to the Associated Press, that never happened: No evidence video color was manipulated in CNN news segment
Here's the original post, taken directly from Instagram:
And here's it side-by-side with what you just posted:
I'm starting to think you need more reliable sources. 😅
splitimageview
Administrator
Feel free to use the AP as a reliable source. 😀 Considering CNN also used the term 'horse dewormer' (wholly false, as Rogan had a medical prescription and wasn't using vet medicine, clearly a pejorative designed to provoke a negative response) this also factors in to the conclusion about the image.
There are many examples of altered images prior to AI, or even using the wrong images (for example, claiming hundreds of thousands of people showing up for an event, using a photo from an entirely different event from even years earlier.)
The Ukraine war was perhaps the beginning of using AI-altered images to construct a narrative.
There are many examples of altered images prior to AI, or even using the wrong images (for example, claiming hundreds of thousands of people showing up for an event, using a photo from an entirely different event from even years earlier.)
The Ukraine war was perhaps the beginning of using AI-altered images to construct a narrative.
Coldkennels
Barnack-toting Brit.
I didn't - I saw the report, found Joe Rogan's actual post, compared it to the Instagram video you posted, and showed with a side-by-side comparison that CNN's video was the real video. The source you posted was the one that wasn't reliable.
"The party told you to reject the evidence of your eyes and ears. It was their final, most essential command."
fub
Established
to be back on topic
splitimageview
Administrator
I was being tongue-in-cheek: "feel free." 😆 Color sure looks altered...but yes, ignore the evidence if you prefer.
splitimageview
Administrator
AI seems to have improved in the last 3 years... 🤡
www.rangefinderforum.com
Ai and the aim to deceive
I did something that I promised that I wouldn't do; argue online about subjective issues. To cut a long story short here is what happened. On a Facebook group about photography, someone posted a picture that the claim that they shot. Looks like the perfect picture, the right place at the right...
www.rangefinderforum.com
boojum
Ignoble Miscreant
AI seems to have improved in the last 3 years... 🤡
Ai and the aim to deceive
I did something that I promised that I wouldn't do; argue online about subjective issues. To cut a long story short here is what happened. On a Facebook group about photography, someone posted a picture that the claim that they shot. Looks like the perfect picture, the right place at the right...
I can assure you that the photo is valid and of thew premiere camera of Lower Slobbovia. It promises to dislodge cameras which have been pre-eminent in the past. Bravo Lower Slobbovia!
peterm1
Veteran
Ai has its place and even now can do a good job in specific niches. AI enhanced post processing of imagery is definitely one of these. I have used a few AI online tools to improve rendering of old images - for example sharpening, colorization etc and while they still have their flaws they often perform much better than alternatives even though there is a bit of "gimcrackery" involved - that is to say it has to take a best guess of how the image was intended to look and make some elements up where it is missing. It's the latter bit that is a bit of a worry. If too much of an image is missing and has to be made up - is the result "real"? That's the question. We already see frankly fake stuff on channels like YouTube where videos containing still images purporting to be historical fact include imagery that is obviously fake to those who are familiar with the subject (but not necessarily to more ignorant souls who take it as real - which is what seems to be intended). I am yet to see any such videos which contain disclaimers that tell the viewer about the made-up stuff contained therein. That might at least be a useful starting point.
As to the more mundane uses of AI in photography processing, the following photo was from an old black and white image of my grand uncle and grand aunt visiting Adelaide from New Zealand probably in the1920's. I used an online AI tool to sharpen it a tad and colorize it but that's all. It handled the tasks reasonably well (though certainly not perfect) without the addition of anything new (or fake) that was not there in the original photo other than colors of course. That is to say the final image does not transcend the levels of post processing that I myself might do "by hand" in such a photo. So, I certainly have no objections to AI used in this manner. It's when stuff is arbitrarily added or subtracted or "morphed" into something entirely new and presented as real life that I begin to feel the "heebie-jeebies" creeping up on me.🙁
Edit: On thinking about it further there is considerable finesse in the choices made by the AI engine. For example. It has chosen what look like pretty accurate skin tones (Though God knows if it could discern someone with black skin for example. It possibly could do so from the tonal values of the skin. The green in the bushes look very realistic - the green in the bush at the right is in sunlight and is represented in lighter tones but also light green than that on the left which is shade. Again it has made a pretty creditable guess about colors. I might say the some of the colors in the grass - some of it looks a little dry and sunburned - exactly as it appears in an Australian summer. The pathway (presumably fine gavel or perhaps concrete looks pretty accurate too, though possibly a little too "cold" considering the bright sunny sky. These is a little too much blue also in the shadow on the man's pants, so maybe the AI is taking its clues from the color assigned to the sky (which, however itself looks pretty good.) I am possibly most impressed though with the almost rusty russet coloured tin roof which is entirely consistent with Australian vernacular homes of that era. Also, the colors of their clothing which to my eyes looks very realistic especially for that era. Is the Ai smart enough to "know" this is from the second decade of the 20th century? Possibly the style of the clothing or of the architecture provide clues? Or is it a "fluke" - a lucky blind guess. I suspect not. Even though some other color choices could have been assigned to clothing and appear broadly acceptable, these choices work pretty well for me not just because the colours look generally real for clothing, but because they also look "historically" accurate. As a kid I recall people of my grandparents generation dressed in pretty much the same style and in similar colours. So, I imagine that if the AI can be trained to associate a style of clothing with a particular era (and gender) it can be trained to associate clothing colours with those styles, eras and genders too.
Last edited:
- Local time
- 11:49 AM
- Joined
- Jun 23, 2005
- Messages
- 6,360
- Local time
- 11:49 AM
- Joined
- Jun 23, 2005
- Messages
- 6,360
By some perhaps, I don't.
Censorship in my eyes is where post, opinions, images, thoughts are removed where there is no publish set of rules as to what is acceptable and what is not. Secrecy, purposeful lack of completeness, specifically gelatinous descriptions and more can all play a hand in the act of censorship. While not communicating how/why offenses have occurred is often a part of censorship, it's not a requirement. I've experienced censorship on several annoying occasions on Reddit, sometimes at from BOTs, some times from liveware. Either way, when I ask for clarification or try to make my case as to why I feel they are mistaken, crickets.
Not a shock, I do not look at anything that has happened at RFF as Censorship. There have been bad actors who have refused to follow clear and concise rules that have been banned. What the SOP is to provide warning and guidance to people who transgress the lines set up over 20 years ago. It's worked, other sites have come and gone, few have stretched as seamlessly and as far beyond where they started and still kept the spirit of expression alive as well.
For me It's a lot more than just that, but that's me. Every mod has their own approach and thoughts. We've been very lucky to have some amazing people step up to help out. Some for longer than others, all contributing to one of the best places on the Net.
We don't work in secret, we strive to help make RFF ever better and on good days, a fun place to read and contribute.
B2 (;->
Dogman
Veteran
Well, I guess we've proved Pan's original point...right.
................................................
................................................
yossi
Well-known
Agreed (but this could jolly well be Ai generated 🤣).
Mos6502
Well-known
A little on-topic levity for this thread. Warning: strong language included.
raydm6
Yay! Cameras! 🙈🙉🙊┌( ಠ_ಠ)┘ [◉"]
Fujifilm is already working on it in their healthcare business segment (medical imaging).
Wonder if this technology will find its way into their electronics/consumer segments:
healthcaresolutions-us.fujifilm.com
Wonder if this technology will find its way into their electronics/consumer segments:
Synapse AI Orchestrator
Get unprecedented insights from Fujifilm's AI Imaging Orchestrator. Better insight turns into better outcomes. Contact us today! Your patients will thank you.
healthcaresolutions-us.fujifilm.com
valdas
Veteran
Healthcare is actually one of the areas where AI benefits can (and already are) very significant, especially in diagnostics, drug discovery…Fujifilm is already working on it in their healthcare business segment (medical imaging).
Wonder if this technology will find its way into their electronics/consumer segments:
Synapse AI Orchestrator
Get unprecedented insights from Fujifilm's AI Imaging Orchestrator. Better insight turns into better outcomes. Contact us today! Your patients will thank you.
- Local time
- 11:49 AM
- Joined
- Jun 23, 2005
- Messages
- 6,360
Wonderful image, at least Photoshopped. But then again his hands look big for what I think they should be.
B2 (;->
Thread viewers
Similar threads
