Mastodawn

@sjn
The use of AI is not relevant for quality. One produces good or bad products with or without AI use.
It is definitely dependent on the human side, whether or not her/his homework is done. Let me say that I saw shitty code produced by humans and AI, as well as good enough code.

@gisgeek I think that strictly within the software development field, you may have a point - under the right circumstances.

Sadly, these tools aren't _only_ used for supporting highly skilled software developers.

Just take a look at your profile photo - clearly generated! What do you think this tells people about yourself?

This is what I'm asking in the poll: Does the next person seeing that image associate it with a positive, negative, or no change in quality?

Makes you think, no?

@sjn
Ah nice example the image. Let me explain. Incidentally, I'm perfectly able to draw a self-portrait of myself in Moebius style. But I had no intention to do that for a series of reason, including the time to dedicate to use ink and colors for that (I'm an old fashioned amateur comic book artist). I deliberately choose to not doing that. So the use of AI says exactly nothing about me (i.e, it is not relevant) which is the point. Did you draw your avatar personally?

@gisgeek My avatar image was drawn by an illustrator on a commission. I don't have the skills to draw. 😅

(That reminds me, I really should reach out to them to commission an update)

Though my question wasn't about your intention with the image, but rather what the audience/reader associates with it, when seeing it.

I think that their thoughts matter, though of course this doesn't have to mean their thoughts matter _to you_. (And that's fine, really. You do you! 😸)

@sjn I understand the point of view of artists and creators. Being used for neural net training is not something many of them have ever contemplated. Which is fine, but licenses and copyright exist for that.
But it's a totally different matter. Again, it is not about quality, and I could cite that photography was not considered art in the old days. At that time, a drawing was art, a photo a mere reproduction of reality. Perceptions of such things change a lot. We live in interesting times.

@sjn It is also one (not the main) reason why none of my drawings have been published on the web by me. Sure, I could add a license and copyright (but I would probably use a CC-BY license), but that would not prevent possible abuses.
Of course, creators are now extremely worried about their role and future, but none did the same when CGI was introduced in cinema (and that is largely computer-generated, with no ridiculous marks).
So maybe people should b more ehm, coherent...

@gisgeek what do you mean with "be more coherent"?

@sjn Simply, I see such AI things as yet other tools; it is not the end of the world, and fighting against them is no different from fighting against cameras, digital art, CGI in cinema, the whole cinema (versus theatre), and so on. So why AI mark only? One could add the Handcrafted mark vs Industrial. Is that a quality marker? Not necessary so: a lot of handcrafted things are simply bad products, plain and clean.

@sjn
Anyway, thanks for your poll; it sparked a possible blog post where I could better articulate why quality is a human-driven goal, not something intrinsically present or absent in AI-aided design. High-quality or good enough are often the choices in many fields, regardless of tools.

Dirk Willrodt Feb 20

@gisgeek @sjn I would go further with the whole quality of the product. For me, the use of AI tells me something about the quality of the user.
I shall update my profile and be more rigorous with blocking AI users.

@gisgeek I get your analogies, though I'm afraid they might be failing you.

The #LLM tools today aren't like CGI or digital art or the introduction of television.

Your examples are are tools of the hand and of trade and media.

#AI tools are tools of the mind. They aren't just a support for thinking, but increasingly a _replacement_ for thinking, and this includes all the consequences that come from this.

We regulated the use of tools to avoid the bad behavior. Maybe do the same with AI?

@gisgeek I think polls like these are useful for identifying where it would make sense to introduce a "Drivers license for #AI".

Clearly, this tool is being used to hurt people today – in too many ways to list here.

Does the few positive/constructive use cases weigh up against the damage that is done by #LLM tools today?

Clearly not.

And this is important because we don't live on our separate software/tech bubbles. We live in a society, together with everyone else.

@sjn Ah sure, the impact of AI on society is a much larger topic, my observation was only about simplicistic labeling of quality for human-only vs AI-aided tasking. I'm worried too for many aspects of the AI-revolution, but which are largely due to our (as a whole society) total incapacity of managing changes in a proper way, for instance in order to avoid leaving people behind.

@gisgeek The impact of #AI on society isn't actually such a large topic.

We can easily cut right through the rhetoric and complexities, and ask one simple question:

Does using #AI help us create a society that is better for all of us?

Those who say yes tend to be of the techno-optimist type, always hoping, looking forward, maybe a bit naïve?

Those who say no tend to be the realist type, looking at what happens today and shaking their head in dissent.

Those who are deep experts, shout #NoAI!

@sjn ROTFL, if the world were in black and white, we would have lived in a perfect society for ages. Unfortunately, there is nothing that is good or bad a priori, and the future is always in the fog. If one were to base a decision on what it seems at the present time, we probably would still live in a forest. Changes are never good for all people, so the logical decision would be no changes at all.

@gisgeek Yeah, nothing's perfect, and expecting something to be so is just as naïve as huffing hopium while waiting for someone else to decide how our future turns out.

We get to ask the questions, say what is desirable and what is not, and use our influence to help steer our society in a positive direction, through political engagement, lawmaking, public discourse and consensus-building.

Yes, this isn't black/white, but simple polls can help us see if _this_ shade of grey is dark or light.

mathew Feb 13

@gisgeek @sjn So you won’t publish your art because you don’t want to see it abused against its copyright license, but you’re fine with using AI that ripped off millions of artists?

Russell Phillips Feb 16

@gisgeek @sjn

licenses and copyright exist for that.

Yes, they do. One of my big frustrations with LLMs is that AI companies violated licenses and copyrights on a vast scale.

Yet, when creators seek recompense for that, we're told that can't be allowed to happen because it would destroy the AI industry.

@rpbook @sjn
Clearly, a lot of training has been conducted in violation of third-party rights. But note that the violation, in most cases, has been recognized not for the digitalization — processing—destroying part, but for the use of a clearly pirated repository of digital content (see the Anthropic case). Like it or not, the training part is not, if not explicitly introduced as an exclusion in the license, a violation.
The same for FOSS code.

@rpbook @sjn
Also, for the GPL use, note that 'derivation' cannot be confused with a set of billions of weights. The key point is the possible use of non-FOSS code in training again. But all that needs to be demonstrated.
Of course, IANAL, but I see very little possibility of seeing such points in a judgment.

@rpbook @sjn
This is, unfortunately, also the main reason the so-called ripping off of artists' creations is pointless. If you buy a book with pictures of original creations, one can use them for training, exactly as a reader can study such portraits for their own goals, make hand copies for their own use, and so on. Like it or not, licenses and copyrights are something more specific than what it seems the idea of many people.

@rpbook @sjn
All that just to say that licenses probably should be reconsidered for modern times, because they are quite inadequate for some people's vision. If you have concerns about the use of such personal creations, let me say clearly: put them in your drawer.

jan Ki | 奇

Feb 18

@gisgeek @rpbook @sjn
pardon my French, but what the fuck are you blabbering?

legality doesn't matter for the question of whether or not you should scrape the web and abuse art for some uncredited heuristic amalgamation without any artistic vision or value

a couple of assholes (including users of LLMs) burning the planet for inherently disrespectful consumerism isn't something that should stop you from still creating and publishing, unless you've already given up on life anyway

Russell Phillips Feb 16

@gisgeek @sjn I'm very aware of the Anthropic case, I'm a part of it.

Part of their defence has been that if they have to pay damages for everything they pirated, they'd go out of business. And now governments are talking about adding AI exceptions to copyright laws.

Telling people to not share things so they don't get stolen is not a solution. It's simple victim blaming.

@rpbook @sjn
The truth is that copyright and licenses cannot be used to avoid abuses per se. They need to be defended in court, and I'm quite sure copyright laws will change, but you know that such laws vary from country to country, so the problem was there before and will be there in the future as well. In the past, changes in law always followed changes in technology. I see no signs of something better for the future.

Russell Phillips Feb 16

@gisgeek @sjn

The truth is that copyright and licenses cannot be used to avoid abuses per se. They need to be defended in court

And when creators try to do that, we're told that the law can't be enforced because it would destroy the AI industry.

They had the money to buy ebooks for training, but they chose to pirate the books instead. Then, having been caught in a clear case of copyright infringement, they're trying to avoid any consequences.

Amira Magdalena Feb 20

@rpbook @gisgeek @sjn laws exist to prevent companies doing bad shit, namely with the threat of them going out of business due to financial punishments, the fact that any court would even tolerate that defence really makes a mockery of the legal system

Corinna (versiffte Göre)Feb 13

@gisgeek @sjn it tells me that you consider your profile picnot worthy of any effort.

Like my all year Christmas version of my photo tells people I'm too lazy to switch my pic accordingly.

Cassandrich Feb 15

@gisgeek @sjn It says a lot about you. That nothing you say is worth anything.

hauch Feb 18

@gisgeek

I wouldn't say, that it says nothing about yourself.
It says what you're willing to show others as a visual representation of yourself and it's not flattering.

That you're willing to use "AI" for your profile picture at all already says a lot.

It reminds me of all those incels who used to have profile pictures of half naked women or their "anime waifu" and didn't get, how this is an issue, especially when approaching women.

The lack of self-awareness is similar.

@sjn @gisgeek "I think that strictly within the software development field, you may have a point - under the right circumstances." hard disagree, and honestly, people thinking that putting bits of already existing code together until it looks like it working is the same as software development is insulting to say the least.

Like, I can heap a lot of actual shit together in a river until stuff can pass to the other side and call it a bridge... but that doesn't make me an engineer.

Salve J. Nilsen Feb 16

@ainmosni @gisgeek Yeah. It's kinda funny too, actually.

How long will it take (and how expensive will it become) before the "Vibers" out there find out software development isn't as easy as they imagined? 😃

🍿

Daniël Franke

@sjn @gisgeek This feels so similar to cryptobros trying to speedrun all the mistakes of a finance system.

Burak Gürsoy Feb 16

@ainmosni @sjn @gisgeek that sounds like being similar to the hype part. For marketing, likely, it is better to hide all of this behind a fine(r) print.

Aron (Sten S.)Feb 20

@ainmosni @sjn @gisgeek

Funny thing is, this isn't even new. I had students in the past (~2008) who blatantly copied and pasted stack overflow solutions to a compiling program that almost did the things they wanted them to do.

Unfortunately these programs also did a lot more, a lot wrong and didn't show structure or a recognizable thought process behind it.

These students couldn't even describe what their code, that they claimed they wrote themselves, was supposed to do.

Feels a lot like vibe coding nowadays, with the main difference, that each if these students realized that they had to put in the effort to learn their basics to actually receive their grades (Most of them did) and it was part of their learning journey.

IMHO this stage of realization is missing with most vibe coders nowadays, so these people never actually start a learning journey.

Ben Ford

@gisgeek @sjn you’re missing the point. The question isn’t whether #AI will help generate better code or not, it’s what effect the presence of a “made with AI” badge would have on perceived quality.

Yes, a skilled programmer can absolutely use AI to generate even better code, for every one of them there are at least ninety-nine other goobers gleefully churning out slop as fast as their slop churning machine will go.

This means that when I see a “made by AI” badge, there’s a 1% chance it’s quality and 99% chance it’s slop.

Matty Feb 13

@gisgeek @sjn mildly agree with this. though humans can make poor products, llm cannot make good ones.
so it is a choice of either 'mediocre to rubbish' from AI, or, 'good to mediocre to rubbish' from a human.

and lets not forget at the end of the day, it is a human pushing the product forward, either human or AI made, and it is a human profitting from it.

but if it is AI made, then the one profitting does not care about the wasted resourcers and possible misuse of intellectual property (...)

Matty Feb 13

@gisgeek @sjn (...) possible misuse of intellectual property.
I will not help said human profit if it is up to me.

in the end, it is not as much as 'made with AI' marks it as poor.
it is that it marks the product as Not good

Aron (Sten S.)Feb 20

@gisgeek @sjn

While you're right with what you say you circumvent the actual question.

It's about the expected/perceived level of quality, not the actual quality (and maybe about morality).

If the prices are the same would you buy the organic fair traded oranges or the ones that come from a company known for exploiting their workers and not caring for environmental impacts of their production?

Leendaal Feb 13

@sjn i mean in the end this is not about the actual result just about a signal ai in general IS error prone because it has to be.

So quality control aside, someone decided to go for an error prone production process and a large quantity of product both these things signal a lower production quality. And no single individual cared enough to do it themselves.

For art, for me art is a human expression, ai automaticallly fails. Can it create pleasing images or the like sure. Doesn't matter.

Bigfood Feb 13

@sjn
I'm not in to art, only listening to music.
I would notice the difference between a AI picture or a human made one, if there are no bad errors like 6 fingers on one hand.

There is a AI band, except mastering everything is done by the AI, trained by humans on contract.
The songs are good enough for me to store on my phone.

As far as I got the LLM stuff, everything depends on the prompt.

ATM LLMs are only a better search engine for me, but only used if normal search isn't getting me the results in searching for.

WooShell Feb 13

@sjn I don't necessarily expect lower quality, but I do at the very least know that creative rights have been violated in creating it, so I would be less inclined to buy/use it.

buherator Feb 13

@sjn Assuming by AI we mean LLMs, this stamp would essentially say "no one cared to think this through".

Giacinto Boccia Feb 13

@sjn
Even it the real quality of the product/service was the same, it generally means that it is less interesting

Reay Feb 13

@sjn @cstross I understand some publishers are now marking AI-made books with a mention/logo(?) inside the cover.

Guess who’s got two thumbs and won’t be buying AI generated books.

Victor W Allen Feb 13

@sjn While my assumption is that AI products will be of lower quality in some (not always obvious) fashion, I think that would not be the reason to avoid such a product.

Provenance matters! An exquisitely cut gemstone with a "blood diamond" tag on it just isn't as appealing as its quality would suggest

ulf Feb 13

@sjn
I’m torn, because although in general I’d assume the “made with AI” suggested lesser quality, I have seen a lot of tat made by humans where quality was not a consideration at all…

diffie :3 (bindiff)Feb 13

@sjn

I have chosen “LOWER”, since many products can have undisclosed or unclear additional requirements.

When it's a product that is bought/acquired/provided with expectation of further interactions with that (computer hardware, an application, both off-line and SaaS etc., etc.), then I can expect further requirements that are frequently not completely clear from the product description – such as need for powerful hardware for local processing (quite painful when the product is just SW without the HW), need to use 3rd-party data processing services or services provided by the manufacturer (may stop being provided in few months/years, privacy-related issues, …) and also many others, including increased environmental load/damage (mostly related to SaaS products, dedicated HW tends to be designed to be ± fine and efficient with this).

Hiisikoloart Feb 13

@sjn
AI stamp means it was made with stolen art, excessively used resources, greed, and money of billionare facists. It is void of all meaning, effort, or human element. It is so deep below anything human made that the closest word that I can think of for that place is "abyss" and even that is too kind.

Anyone with a soul and human desency should avoid it and trash that slop the second they see it. Support human artists - always.

Matty Feb 13

@sjn heck, if the made with AI product is cheaper, I'm still getting the other one.

just like I get books from bookstore.org even when they are more expensive than amazon's

Noisecolor Feb 13

@sjn
That's a silly and pointless poll. Of course it's not the same quality. It's like asking if people think a 5k€ suit worked on by a tailor for two weeks is the same, better or worse quality than a 50€ from HM .
What do you think?

Point of automatization is producing products at speed and price for the quality that is good enough. Not quality alone.

Klaus Stein Feb 13

@sjn

This is multi-dimensional. Quality itself is not one-dimensional and “higher” quality is not the only reason to choose what to buy.

Than AI is much more than using generative LLMs. For example cancer detection with machine learning based image evaluation has very high success rates, so I would very much follow its advise (if used by a domain expert). And even generative AI used by an expert as a tool can be great. On the other hand I may want to pay an artist but not an AI company for some image of same “quality”.

And then there is: if an AI can produce it its value will not increase in time while it is from a famous artist it may (see →Benjamin, aura).

There is all the ethical aspects. And…

And this only got me started.

Michael Harley Feb 13

@sjn gosh I think there's not nearly enough nuance here and everybody is going to assume the absolute worst or best scenario in their head.

Did a developer, using Claude in their IDE, carefully guide it to build something, with attention to detail and corrections?

Or are we just talking about AI slop, where someone who doesn't really know what they're doing told Claude to build something and whatever was spit out is what they got and they had not clue or idea about development?

I think AI can make good devs better. I think lazy devs will still make shit. IMO

Evgeniy Latuhin Feb 13

@sjn which one would I buy?
The quality of the product do not equate to using AI or not using it.
But using AI equate to labor exploitation on the bigger scale. And quite probably signal about other unethical practices of the people in the management (or other) chains of the product.
So given the choice I would spend my money on the one without the sticker.

trianderror Feb 14

@sjn
“When AI is mentioned, it tends to lower emotional trust, which in turn decreases purchase intentions,” he said. [...]
“We tested the effect across eight different product and service categories, and the results were all the same: it’s a disadvantage to include those kinds of terms in the product descriptions,” Cicek said.
🤔
https://news.wsu.edu/press-release/2024/07/30/using-the-term-artificial-intelligence-in-product-descriptions-reduces-purchase-intentions/

Using the term ‘artificial intelligence’ in product descriptions reduces purchase intentions

Companies may unintentionally hurt their sales by including the words “artificial intelligence” when describing their offerings that use the technology, according to a study led by WSU researchers.

WSU Insider

Brooks Smith Feb 14

@sjn Not a universal rule, but common enough, IMHO: The people and companies that *talk about* using AI are the ones who are over-using it and are far more likely to produce low quality products.

AI (both the broader universe of AI/ML and specifically LLMs) very much have good and valuable use cases. But they're tools, not panaceas. No one writes "made with Excel" on products even though many (most?) are to at least some degree - because it's merely one of many tools in the toolbox.

Woozle Hypertwin Feb 14

@sjn I put "no difference" because it would depend a lot on the context and how I'm evaluating "quality" -- but I think in today's environment and in most contexts, I would tend to be significantly more leery of something where the maker thinks "made with AI" is a selling-point. If it was more, say, honesty in advertising (e.g. a future where this is a required disclosure), then my evaluation would depend much more on other factors (though for now, it's still a flag against).

olof hennig Feb 14

@sjn i wouldn’t necessarily say lower quality as much as ”if you can’t bother putting an effort in making this, why would I bother paying attention?”

Michael Richardson Feb 15

@sjn The one with the "Made with #AI" mark has no copyright, so you can just make as many copies as you like. It has no value, thus any price on it is nonsense.