AI images: hide. Yeah, sure.
I guess old printed encyclopedias are much MUCH better than this shit for your child's homework. Ugh, this is so disturbing.
(ノಠ益ಠ)ノ I∀
That's because there's no simple way to scan billions of images to determine their provenance.
Even glancing at your screenshot what's to say a lot of those aren't older artistic digital images?
This can be a fun learning example for you and the kid both. Download some massive packs of images online. Millions. Lump all of them into one single folder. Now sort them for common features, which were made with cameras, which were made digitally by a person, which are AI, etc.
Just go to a source and search there directly instead of using a search engine. AI has made search engines annoying as fuck lately. Wikipedia
Yeah, that should be how it's done. Must teach that to children though, this is so brain-hurting.
We shouldn't have to, but yeah... Here we are 🤷
I feel like AI is creating a renewed old way of browsing. Find info about some project in a forum(lemmy) go to the actual website, not social media, not google just plain old type the URL and go directly to a human made site.
Distrust in corpo and AI might just bring back websites made by humans. Personal blogs, forums, product pages etc...
Don't know how they do it, but it is definitely worth your time to mass report these images. Maybe take about 10 minutes and just repirt at many as you can. That's what I do. Probably doesn't do much, but if enough images are flagged, I'd hope they take a look at the website and rule whether to block that site completely. It's what I would do if I were them.
Edit:
Saw in a reply that you have reported them. I personally think we should strive to create a society where genAI is consistently mass reported as what it is. And that starts with people like us.
To be fair, it seems a lot of these images come from the same 2-3 sources. They just pollute the internet with their slop. I believe they have many other slop images for everything so blocking them would probably cut a lot of these results. They might not want to completely remove these sources because they actually have useful images as well but at this point DDG should just cut them out because they're heavily infected.
If this goes on, we probably end up with a curated human-centric small and limited internet at some point and it will be hard to protect it. That probably also means far from ads and SEO-based internet too, since that's what started the pollution in the first place.
Wasn't Google actively pushing SEO, too? Like filtering information pollution wasn't a huge part of a search engine's job.
They were, and initially it was supposed to be a good thing (like every other thing Google takes on and going forward). They made it kinda mandatory if you want to be seen on Google. After Google started pushing ad business too, it became "ads inside SEO". It was not an optimization anymore.
Labels should actually mean something.
Yeah, right out of the gate, we needed to make all diffusion do some kind of tell. Even if we started it today, it would be too far gone.
Our only hope at this point is that good diffusion is so expensive that when the bubble bursts, no one is going to go to the trouble of making it.
True, but standards only work if people actually adopt them.
Fucking hell first row last one 🫠
Presumably DDG's detection has a hard time keeping up with AI development. You can report AI generated images via the three dot menu so that's... something.
Already reported but not sure if we can fight this. We need AI to block AI apparently.
Sorry, Earth. For what we've ~~AI overlords~~ done to you.
DuckDuckGo did a site-wide poll about 6 months ago asking their users if they wanted AI or not. It was on their homepage. 90% answered "no".
What did they do in response? ... nothing. It didn't fit their preconceived notions and apparent agenda, so they've just completely ignored it. The customer is always right, except when they don't want AI, then they're just wrong.
that poll was an ad for their noai service. nothing more.
And now they have AI, despite literally no one wanting it
Apparently, the "hide AI" only blocks some domains that are known to be generators of the images. I did a "waterbear" search and noticed that a number of the slop is on stock.adobe and storage.googleapis, which immediately makes them "clear" for the search
I still hate DDG for leaving the "show ai: yes" as the fucking default. Fuck them
I tried the search on both Ecosia and DDG using the HUGE AI Blocklist (linked elsewhere on this thread) and Ecosia seems to filter out the stock.adobe and googleapis images whereas DDG still lets them through.
What I want is a filter that can rid me of LLM-generated text on websites. I HATE looking for Linux-related stuff and getting that "high schooler giving a presentation" formatted crap that opens up telling me about why I'm seeking the information I've clicked the link for and then summarizes it all at the end.
EDIT: As I scroll a little further on Ecosia I see more obvious "AI" generated images... oh well. At least DDG lets you flag them.
At least DDG lets you flag them.
Part of me thinks that's a ruse for micro$oft to train models, "oh, this one was too obvious, avoid making an image that looks like this!", what with ddg running bing
Yeah sadly the No AI filter isn't 100% accurate, it does try though and is better than no filter IMO
It probably is better, but the sources get worse everyday. Terrible.
Set search date prior to 2019.
That's a sound advice but sadly don't have that option on Images search.
It probably only filters for the machine readable markings that will be required in the EU starting next month.
But as soon as someone crops it in Photoshop it is indistinguishable from digital art made by humans
Probably not in recent Photoshop versions, they are working on preserving some kinds of metadata/watermarks. But an uncountable number of other ways to edit a picture (including screenshots)? Very likely removes both metadata and steganographic attribution.
Good job Adobe (upper right corner) for selling that super scientifically correct stock photo
Bottom center too. Beaver teeth?!
I was doing some arts and crafta with the girls, and i decided to use my photo printer to print them out some puctures of animals, wince they love animals. One of them said i want a funny bunny. So i searched for funny bunny. Holy shit, the result has absolutely disgusting to look at. Just animals were okay, but all it spat out was some boomer ass ai slop of "animals" i hate what the internet has become.
Here's a free real bunny picture I took
Hair on the back: Clearly too long Right ear: Weirdly shaped (AI hallucination) Towel in front: Looks very real but the texture is a dead giveaway The Rabbit: Very cute, ignore everything above because this is just a copypasta.
Yeah, it's so so disturbing. Straight away nightmare fuel. It's so bad that a lot of these sources are auto-made by LLMs. And it's getting worse and worse. :(
I'm suspicious of the whole "noai" thing, especially when it is championed by companies with presence in AI. What wouldn't these companies just capture the noai search revenue to overshadow alternatives, continue to invest all their profits into AI, then kill noai and go full AI when the ideologically pure alternatives have fallen?
What alternatives would that be?
I constantly search images and I can't find anything that consistently blocks AI. I've had better luck finding images through Yandex. Even regular web searches are garbage anymore.
I did try Kagi, but despite the raving from people I felt the improvement wasn't much.
Would love suggestions. I like privacy, but if I can't find shit, it's meaningless.
In this case, you’re seeing “artistic interpretations” based on tardigrades? They do this with exoplanets too.
Recently did an image search on regular DDG, no AI filter, and was astonished by how much AI slop I got. Like, at least half the results were slop. The filter isn't great, but it's better than nothing.
For legit and usable (CC licensed) images, I stick with Wikimedia Commons. Haven't seen any horrors yet.
I'm not sure I understand, are you thinking they are all AI? At least some of them are not - I recognize them from the before times. Tardigrade images have always been goofy looking.
The adobe stock images are AI though. It's probably a safe bet to avoid stock image sites although Alamy seems to have some good ones. Adding the photography filter helps a bit too, but I still got some stock images used on other websites.
Of course not, but a lot of them are, which is not good enough for No AI search. Yeah, stock images got worse, they shouldn't be called "stock" anymore. We almost got to the dead internet part and it came so fast, it's mind boggling.
About the death of the internet, I have slowly stopped casually searching stuff. I almost exclusively search within known domains now :(
There's 21 images. Cautiously I'd identify at least 10 of them as AI generated. The issue is that there are any AT ALL with DDG's options set to hide AI generated images. That setting used to work really well.
Ironically, AI can tell me the adobe images are AI and why they don't get flagged in image search:
Yes, this image is AI-generated.
While checking the image file shows it was not made using Google AI tools specifically, it bears the official "Adobe Stock" watermark directly over an image ID (961863094) that belongs to Adobe's generative AI collection.
When downloaded directly from the official Adobe Stock platform, the full-resolution file contains embedded metadata acknowledging it as AI content, but this background data is lost when saving or extracting the preview image.
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.