While AI images in a blog signify low effort high volume posting where the text likely was just copied from AI without any care.
Just go full American Psycho eggshell business card.
Why these super-pro-AI people take this as such a personal slight is strange to me. You didn't care enough to actually make something, what are you offended if I tell you I don't really want to look at it?
And the tools are going to get better too.
Your comment is an example of what the parent comment is discussing; the reflexive "you can't really tell and anyway you can do it better to stop people being able to tell and also soon it will be better so no one will be able to tell" is a gallop like the narcissist's prayer.
People are telling you they can tell, and that they do care. The toupee fallacy works because there are good wigs that no one can identify unless told, but the idea that that there's actually brilliant AI prose (that no one has detected) always lacks examples.
> "People are telling you they can tell, and that they do care."
In the same spirit, one could argue that "AI Slop" is actually constructive feedback to a small degree. It's a short-hand for "this is sloppy/low-effort/unskilled work with obvious utilization of generative AI tools". It's not much different than using the same photography filter for every post on Instagram or picking the most popular Wordpress theme --- if you are expecting praise for that then you will be disappointed.
The kids are alright.
I mean you're rather boldly skipping over the part where the former is a proven phenomenon (though I would say it's more nuanced than true or false but there's too much to unpack there) versus something that, as yet, I have not experienced.
> You can still want human music for the human reason, but you won't be able to tell them apart.
I will pepsi challenge any media you like, from the most powerful models you like, happily.
And I'm only willing to even entertain images in terms of said challenge. The music and video are so fucking obvious, just absolutely glaringly AI when they are that I don't feel the need to even dignify it with a test. As yet, I've seen some images that certainly get close, but there are always tells in my experience.
You make it a personal slight when you espouse, that your aesthetic preferences are universal and then farther malign people with different preferences by saying that they are not aesthecticly minded.
Good thing I didn't say that then.
> You make it a personal slight when you espouse, that your aesthetic preferences are universal and then farther malign people with different preferences by saying that they are not aesthecticly minded.
Aesthetics are just taste and taste is subjective and will vary by person. That said you can, broadly, achieve "good" or "bad" aesthetics based on opinions at scale. The predominant opinion right now is that AI art is bad. You feel what you feel about that, but numerous prominent people, numerous companies who provide services, etc. are all responding to the fact that people find AI art ugly in a variety of ways. Any stock photo site I use has an option to filter AI art, and if it doesn't I don't use it. Projects in the public space that are outed for using it get backlash. Various social media sites are demonetizing generated content.
And to be clear, me saying people are not aesthetically minded is not meant as an insult. Plenty of people I respect do not care how shit looks or feels. But they are a minority in my experience, and frankly, anyone working in an even remotely creative field probably does care. YMMV.
Be that as it may, though, aesthetics do matter to a lot of people, even if they don't discuss it openly or lack the vocabulary to do so. People will say it just "feels wrong," or it's "soulless." It seems to activate our brains very similarly to things deep in the Uncanny Valley, where it looks allllmoooost right, but then fumbles something subtle in a way that makes it off-putting in retrospect.
I don't think we'll ever get rid of AI entirely. I think it will remain in the margins as cheap solution-oriented ways to make graphics, and maybe video again if they can ever figure out how to make it profitable. But I have a feeling an AI video commercial would run on TV in a similar time-slot to recycled Billy Mays ads, and with similar function: cheap, already there, get the job done, and the audience is not discriminating. It's "fine" in the broadest sense of the word.
Some of my favorite examples:
This stuff you say it signifies, it says more about you than them.
People will always make snap judgments, you can't inspect every website in detail. So if you use AI, make sure there are some other signals of quality.
So now we have this low-effort image generation tool. At least it doesn’t indicate that the author wasted a substantial amount of time looking at pictures. Still a silly waste of time, but at least it is a small one.
A lot of traditional news outlets have a style rule that every article must have an image. Look at https://www.bbc.co.uk/news https://www.nytimes.com/ https://www.economist.com etc - every article has an image. Even a report on something highly abstract - "BP reports record profits" - will be illustrated with a photograph of some BP-branded fuel pumps. Stories on things like cybersecurity and inflation get generic images like snippets of blurry computer code, banknotes etc.
Some bloggers feel their blog is improved by adopting the style of professional journalism, and (among other things) put an image on every article. Even if for a programming blog good lead images are hard to come by.
Interestingly, professional journalism organisations often have pretty strict codes of practice about images, and photoshop use. They have rules on everything from illustrating one rocket launch with a photo of another to posing the subjects to colour-correcting images to creating multi-image collages. So I don't imagine we'll see many AI-generated images in traditional news outlets any time soon.
I seem to recall many instances of the use of incorrect photos, and the earliest I heard of was in the The Guardian in the 1980s before the general decline in media quality and standards.
It's awfully tempting for a war photojournalist to pose the group of tearful women in front of the bombed-out building across town. Or for the celeb gossip reporter to collage together two photos without making it clear it's a collage. Or for the person colour-correcting a photo of a divisive figure to lighten or darken their skin.
And the resulting rules are complex and at times judgement-based, because posing subjects is allowed in some circumstances, collages are sometimes allowed, some colour adjustment is allowed, and so on.
Still, serious journalists aren't going to be illustrating their articles about US-Iran negotiations with chatgpt any time soon.
No-one knows exactly how said images ended up getting this sort of usage. Presumably some random journalist was on Google Images one day, come across pics that looked vaguely like what they remembered the games to look like, and just kinda went with it.
I also remember quite a few cases where fan-made box art mockups got accidentally used as the pictures of the real thing, and when the BBC used a logo from Halo to represent a real-life organisation.
Same goes for book covers. You probably need something unless you want to look really generic and cheap. But I'm mostly not going to pay a designer hundreds+ of dollars for a cover of an independently published book. I have had a former colleague who is actually good with this sort of thing do it in the past. I've also done a crappy job myself, done it in conjunction with a publisher (for whom I had the supply the cover image), and, yes, used generative AI/Photoshop.
It’s flooded the market to the point average users are starting to turn away. Look different or be relegated to obscurity.