Feedback about the inclusion of New Fast line art feature

I wonder how people would react to this feature if it were implemented with a classic algorithm instead of a basic neural network.

A lot of technology people use every day (probably without even knowing) is used unethically (encryption used for malware), for warfare (GPS) and to remove people’s jobs (mail). Yet we acknowledge that it can also be used in a good way (or people just have stopped caring, I don’t know) and that the tech isn’t simply evil just because it’s also used for evil. I’m against pretty much all AI stuff that exists currently. Yet I wouldn’t say the technology itself is evil per se.

7 Likes

One frustrating aspect of this debate is how many are neatly categorizing artists into ranks of good and bad – talented and not talented, skilled and not skilled – to then use that hierarchy to dismiss arguments instead of considering them on their merits.

The value of your words doesn’t rest on your ability of wowig people with your work. A beginner’s word isn’t worthless because they didn’t reach the point they want yet. Talent? As in innate talent? It’s an illusion! It’s so rare you might have never ever met any artist like this. 99.999% of the artists you look up to sit on a mountain of bad drawings that’ll never see the light of the day.

You’re not your work. You aren’t a failure for struggling, morally inferior, or devoid of value as an artist and as a person. You don’t sit above other people either when your work stuns people, you’re not more valuable, with greater rights and entitled to be listened to and to command the masses of “lesser” artists. Disagreeing with you isn’t a defect of character.

The only thing that someone’s “bad” work tells about them is that they didn’t have enough time and opportunity to reach their goals. It doesn’t even tell us if they’re beginners though we often use it as a gentle shorthand for it. It may be that they simply didn’t find a way yet to study and practice art that accomodates their physical limitations and speaks to their minds. We’re all on a learning curve.

What being skilled tells us is that the person is likely a long time at it. For being a long time at it they have insight into the process, and perhaps into the industry when they also practice art as a profession. Only this. It doesn’t give authority over the subject or worse, how other people should work. The idea of decreeing how other artists should work is absurd. What the hell, people?

The same is valid for being a professional or hobbyist. The point of weighing the words of a pro is not that they have clout because people throw money at them for their skill, it’s because it affects their livelihoods and to try to exert control over how they work is not only to try to control how their art is made and looks, it’s to deeply impact their ability to survive.

So should disabled artists points be listened to with extra care for similar reasons. It should go without a saying that what isn’t a problem for you is an obstacle to others created by no personal failure of theirs.

So, can we please not weigh arguments on the “betters” and the “lessers”, the ones doing rightart and wrongart, and consider instead who will be affected how by this proposal?


Another point is potential bullying caused by this feature.

Yes, it’s worrying. I don’t mean to minimize or brush aside this concern when I say it though: Someone set on hating you will hate you no matter how hard you contort yourself into knots to shield yourself. They’ll hate you your work your cat and the way you eat crackers.

Nothing you can ever do will be enough to change their mind. This isn’t an issue Krita can cause or solve, this is a societal issue.

They’ll accuse you of being a “lesser” artist and point out at Krita having “ai” features as a proof, and when they lack anything to misrepresent as proof they’ll simply make it up. Thing is, does this hatred have any concrete consequences to your life?

Having large groups of people sicced on you certainly has, having your professional reputation has, but thankfully most bullies are just john2903781 from the internet. Chuckle a little at the absurdity of them proclaiming they know who you are from half a dozen words and a 12am sketch, then block and they’ll forever disappear from your life. You don’t owe them your time, energy or attention they clearly crave. You don’t owe them any proof. Boy, they wish. :joy:

Don’t take it as “grow a thicker skin”, that’s horrible backhanded advice – as if resilience grew on trees and lacking it was (again) a character defect! Take it as a reminder that if you step back a little you can find ways to protect yourself that don’t involve constantly worrying about not providing ammunition to people out to hurt you, because that anxiety alone, that stone around your neck, it will consume you.

And in the worst case scenario, if you’re set on having a so called clean copy of Krita as proof that you don’t rely on any aid for your linework – then you can always fork the project and share the fork so it’s verifiable by any finger pointers.

13 Likes

@raghukamath I’m generally referring to ai, and this fantasy about things I’m saying needs to stop. I repeat, I no longer use krita, so don’t need to fork anything, and doubling down on the fork it/‘get lost’ is only confirming every doubt I have regarding krita, and backing up every point I’m making. I didn’t come at you, but, again, there’s unpleasantness no matter how reasonably someone expresses valid concerns about ai/intel/attitude to users. You register that I have strong determination; many do against ai, and it would be helpful generally if those pro-ai could realise that this determination won’t be worn down or put off course.

@ceres you made some good points, and it’s really nice to see someone considering the disabled also; very thoughtful. And I agree, some blindly hate, no matter how positive a person is, but if the person in question has a clean conscience, and understands clearly what’s going on, being honest as necessary then paying no heed I’ve found can work well. It’s never about letting any negativity dent the day.

That’s something I’ve asked 2 days ago, but I didn’t got any answer :upside_down_face:
Not sure if it’s the tool itself, the technology behind the tool, or the user using the tool that is a problem.

And I’m wondering what will happen if some existing tools are improved with some AI technology (resizing with pretty good upscale feature, filling with amazing gap management, miscellaneous GMI’C filters, …)

An AI that can help research in any domain (astronomy, mathematics, medicine, physics, biology, …) can’t be ethical?

An AI that can help disabled people in daily task can’t be ethical?

The day when Parkinson’s disease will prevent me from holding a stylus to draw lines properly, I’m not allowed to have a tool that can help me because some people consider is not ethical to have it?

Problem is not the AI itself.
Problem is how the AI is trained, who use the AI, and what is done with the AI.

I’m against AI that are trained on stolen data.
I’m against AI that are used to scam and flood market with a huge amount of artworks created from nothing else than a prompt or from a copyrighted content.
I’m against centralized AI services that is now everywhere: Apple, Google, Microsoft, Facebook and that start to be imposed.

But I don’t have anything against the AI itself.
And if some people use it in a smart, moderate and ethic way, it could even be a good thing.

So Krita’s developers must be online on this forum 24h per day, 7 days a week, and react to all topics? :sweat_smile:

This topic started to be hot during the week-end.

  1. let the developers having a break the week-end, they’re just humans like us
  2. let the developers read everything: with all the content written here, it’s some hours of reading
  3. let the developers get a talk on their side about the situation and let them time to provide a proper answer

You can be sure there’ll be answer(s) from their side, may be not today neither this week, I don’t know.
Just let them time.

Grum999

6 Likes

You are not at at all listening to us and are implying that the feature that is going to be implemented will be bad, even though we are repeatedly telling you it is not like other generative AI.You are blindly making assumptions. Your concerns were listened and we (the forum members here not the devs) told you that it is not the same. So why are you jumping to conclusions.

On another note let us see what assumptions and false statements you made. I am quoting some text and calling out falsehoods from your first post -

This is false. kuserfeedback is off by default officially. If it is on ask your distro why it is. Also they do not get any meaningful information and the information they get is anonymized. You know it comes under GDPR and KDE is serious about this. You can see this forum too we do not have any trackers here. The only trackers you will see is from youtube videos that people post here. This is not even in krita why bring this up here.

Can’t say anything without context. We have a massive thread talking about AI. Where people like me have expressed grave concern over unethical AI and usage of artwork without artist consent.

This is false. It is not nearly ready it is in development and it is announced without any malicious intent of hiding like your implying here.

Because those are closed source and they do use the very model that you despise. If someone implements a plugin or feature making use of these feature I am sure people will not say no but for that these things needs to be compatible with Krita’s license.

Again making assumptions and spreading FUD. making false assumptions on bad faith and accusing the team of malice done for money.

I have given many example of your bad faith if you really want to join the discussion stop accusing people and put forth your point without making wrong and false assumptions. These are literally symptoms of spreading misinformation by making false assumptions. I have given enough benefit of doubt to you but you seem to be here for spreading false information based on your own false assumptions.

10 Likes

I’ll give my 2 cents as an outsider (Blender artist here, but I’ve used Krita on occasion), so take this with a grain of salt. Based on what I’ve read, I don’t see any reason to be concerned about this feature.

To me, this would be the Krita equivilent of Blender’s AI denoising (which is actually from Intel and Nvidia, whereas this new feature I presume will be built and trained in-house), added back in 2019; it’s a feature that still requires a good base to get good results, but it allows for faster iteration. To say this is like GenAI would be wildly inaccurate.

I’m also not concerned about the Intel sponsorship. It’s worked extremely well for Blender for a long time, I can’t see why it would be an issue for Krita now, especially as Krita is part of KDE, which has a great track record (though I might be biased as a Plasma user c:), and it helps improve open source software.

Again, take this with a grain of salt, as I’m speaking from an outsider perspective, but that’s just my 2 cents

12 Likes

Thank you for the inside.

This is exactly what everyone needs to hear. AI can be useful and can be ethical too and blender proves it.

@raghukamath I think the phobia is people hearing AI and not caring at all how and what it is achieved. Best solution would be to rename to like auto line art generator. At least this will cater to people who are not in this forum and and have this AI phobia

@Grum999 Commendable to be against stolen data, false artwork, and centralized ai services that big tech is trying to impose, for sure. That’s awesome. I think we have a different view about whether ai can be ethical, as I don’t believe it can (have explained earlier). Do pro-ai supporters not realise that factually ai also includes much worse csam? And the accompanying worse censorship of people who would never be involved in such things, as you refer to regarding big tech’s push to impose ai. If there was ever a reason to unplug ai right now, worse csam is it. Putting any flag, no matter how small, on the ai side, does add support to ai’s overall activities.
I’m not sure where you get that I was blowing a whistle for instant roll call to summon the krita team genie-style, lol. I was referencing observing no comment from krita to users’ concerns, in the whole thread, which suddenly struck me as odd, plus raghukamath had said a day or so ago that he was paging them.

@raghukamath I’m not wishing to continue this with you. There’s too much in your post to disagree factually with, so I’ll leave you with your thoughts. I’ve already fully explained every point that you’ve brought back up e.g. I answered that I don’t use glaze/nightshade, don’t need it as I don’t put art online, and don’t want closed source, but was wondering if something could be done that’s gpl/libre. If there’d been a wish to discuss things positively, I’d continue talking with you, but it’s unfortunately not bringing any solutions that can help about users’ concerns.

@abhifx let’s hope ai doesn’t get worse, to the point that people regret calling those who warned about it ‘ai phobic’. That’s ludicrous. Are the AAA artists taking legal action against ai companies ‘ai phobic’? Are those who’ve lost their jobs due to ai ‘ai phobic’? No. Facts are facts.

Wow…

Let’s just take a deep breath and count to 10. I heard that helps to calm down…

No one ever said they are for stolen artwork and false artwork. We are for models, which are algorithms that got better over many iterations, that help to speed up tasks, which we deem take a lot of time, e.a line art, base coloring, etc. Why should we “waste” time with things we have to do over and over again?

I don’t know how you are interpreting AI… I understand it as: give input, and you get a processed output. And that output would be similar to what a human intellect would have done, only that it does not take that much time. To me, it sounds like a lot of timesaving. What about you?

The problem of the big tech ai is how it is processed or how they got the model. They took our data (most of the time with our consent as we had no other choice if we wanted to use their services) and trained their model. In the case of image generation, it was even worse, as they took the images without permission.

In the case of this lineart tool, the krita devs ask for donations. They will give credits to all donations, and if there is stolen art in those donations, they can exclude those and retrain the model if it is necessary. So what is the problem you are afraid of?

Artists who don’t learn how to draw line art? We talked about it.
Stealing someone’s sketches and publishing the finished pieces as their own? We talked about it.
Making the tools for those thieves easily available? We talked about it.
Did I miss something? Either way, we talked about it.

We did not dissmiss any of those concerns. We talked about it and gave our point of view. Everyone can now make their own judgment about this tool.

In my opinion, it is a valuable feature, which will later on be the base of more advanced features, as when the lineart-tool is created, a fill-tool can be implemented, which uses the generated lines as boundary. So no more missing pixels in those flat colors. How cool that would be. Or from sketch directly to color. So much time saving.

Please note:
There will be some people who will use tools and make copyright infringements. But those people already use free and open source projects to steal art. If this tool is implemented or not. Try to stop me from taking your collage you posted earlier. I could paint over each and every image and do whatever I want with it and publish it as my “own”. I have that knowledge and the tools available to me. And it has nothing to do with this new tool. Don’t worry I will not, but you get the point.

1 Like

Dear @elevate

I tried not to directly reply to you but here we go.

I think all of us here agree the AI spread in common parlance is bad. but why is it bad?
Because it takes other artist work and then use that style to generate images by people who are not practised by artists or just want to make art without thinking this is done on stolen work right? It sometimes directly throws the original art.

I am an artist and I hate this too.

However, in krita with this feature, we are not going to crawl on the internet, use artists work and generate line art. Instead, some of us will volunteer in the name of open source and CC license so that it can be trained. Thus making the tool robust.
Accordingly, kritas AI usage is very ethical (in my eyes) than the big corporates that behind close wall just use anything that gets in their hand.

If this doesn’t still calm you down then I am afraid there is a genuine phobia just in the name of technology. and you are definitely free to use any other software which you deem is best for you.

I will be the first one to leave krita if AI is in krita goes the other way but clearly the implementation is in true open source spirit. (I am a long term open source enthusiast).

4 Likes

I quoted your false statements. You should retract those statements and do not spread falsehood.

Read my message again I have included your own lines of false statements.

You are lumping us with corporates who steal data and do unethical things. You ignore our answers and generalise things. For you the word AI means bad. You are parroting the same lines again and again.

We are talking about the feature in krita and not general AI. This thread is about line art feature in krita not stable diffusion or mid journey. For that there is another thread.

In any case I will not waste my time and I suggest everyone else also to ignore this. You are not able to comprehend what we are saying. You do not even use krita now so there is no need to convince you. You can form any opinion just do not spread false information.

Also I did not tell that I am paging anyone. That statement is also false.

6 Likes

Are you trying to say that the fast line art tool is evil because some other model of some unrelated AI was trained on child porn? And since it’s roughly the same tech it is just the same as having trained on such material? If not, why even mention it. If that’s the logic you apply I am really wondering how you can use the internet at all. Guilty by association, every use of a technology is the same as the worst things done with it. This is getting out of hand here, I’m removing myself from this thread if this is the level on which arguments are made. Next the line art tool is going to be called racist because large language models have been proven to be quite racist in the output they create. Although that bears no relevance to the feature we were supposed to be discussing here since it is not a generative LLM. You made some good points I supported in the thread but now it’s getting crazy.

4 Likes

Hello, just a little message to say I support Krita’s team because they explained clearly what they try to achieve and how they will do. I don’t see any threat against “real artists” and the ability to create a fork if you disagree with the politics of Krita is the best answer to give to people complaining. Make a fork. And do art an IA will never be able to simulate, it will be great art.

7 Likes

@elevate
The only downside I can think of is it can help normalizing the use of AI technology in general which is currently being led by big techs, because people tend to just perceive it as the same AI usage. But it can probably be hashed out as more people try to distinguish ethical usage of AI from the current generative AI practice.

Ah and there’s the erosion of ‘what human is’ by the tech itself if you zoom out the scope to the massive scale. If that’s what you’re ultimately referring to by any chance, I can kinda see your concern that AI cannot be ethical in a long(like, very long) run. But I’m not really worried about it because I don’t think it can be diverted anyway.

1 Like

Not sure if that holds too, for example krita allows Gaussian blur or motion blur using Algorithm. If I want to believe “what human is” then I should be able to paint those effects by hand rather than using the tool.

The whole digital painting would fall in that rabbit hole.

On the other note not related to reply to @acc4
I think this thread has gone in a different direction, we were supposed to discuss the nitty gritty of line art tool instead of glorifying “AI in krita is satan in disguise”.

3 Likes

@BeARToys I was responding to Grum’s post; check that, which will clear things up for you. And I’m not ‘afraid’ of anything … posted out of concern, to join in with others expressing concerns also.

@abhifx I’m not sure it’s being understood … krita has gone from ‘we’re against artstation/ai’ posts on twitter, to bringing in something ai, with intel money. Quite a change. People have already posted about being surprised at that, and not able to trust or recommend krita any more. When a change like that happens, it’s natural to be concerned if more ai would come in. People expressing concerns is to try to remedy this, but it seems other people want things to rot and get worse rather than avoid that happening.
And, for the final time, not ‘phobic’ or in need of being calmed down.

@raghukamath … encouraging others to ignore me now? saying I’m a liar, lol, and that my views are a waste of time? I’m surprised I have to type this, but it’s breaking forum rules, and keeps on confirming any doubts I had about this feature coming into krita, and what more could come from that.

@Takiro I’m saying if you plant your feet in the ai camp that then gives support to the ai camp, which includes all kinds of bad things. Obviously my concern is about those bad things, which impact on people. If you choose to see that as me being the devil incarnate, that’s surreal.

@acc4 thank you for your reasoned response. Yes, it’s the normalising, planting feet in the ai camp, no matter what impact that will have on people and what matters to people. The ai lineart tool is doing that, especially judging by the general tone of this thread.

Sorry.

As I said, I think it can probably be hashed out as more people try to create an ethical camp of AI, different from the current generative AI camp. But if you don’t believe there can ever be such a thing as ethical usage of neural network, it’s probably impossible to find a common ground on the issue of this feature.

Anyway, I’m out of this thread now.

Yes you are a liar and I have given quotes from your first post which are lies, first go and correct those lies. As simple as that. And if you do not stop spreading misinformation and FUD then we will be forced to silence you.

Check the list of lies you spread here - Feedback about the inclusion of New Fast line art feature - #105 by raghukamath

4 Likes

I think what would help is the krita team draw the line in the sand somewhere. This with regards to AI is as far as it goes, not a pixel or grain further.

Anyways, as long as if I never use it, the feature is not once being utilized at all, I’m okay with this. It’s quite frankly a tool. If you never ever use it, it won’t have any impact on your work in any way shape or form. It won’t be monitoring what you draw and then feed it into a database, it has to be explicit opt in, NOT opt out. That’s why this is okay though I will admit that the line in the sand needs to be drawn somewhere to assure everyone who is reacting harshly can settle down.

2 Likes