#killer-laurent

for example those shots you posted of his older work with slightly different style.. especially the one in the middle with Les, Louis, Marius and Armand together.. it's fantastic and you can tell they have real talent. and there's nothing wrong with tracing why are you so butthurt by tracing its how people learn…

I always thought their rendering was exceptional whenever it popped up on my screen, but it didn’t strike me as being obvious AI, at least not compared to other stuff I’ve seen, where people have obviously just taken an existing picture and have AI slap an oil painting/watercolour/whatever effect on it. Sure, the vibrancy of colours and near luminescence felt almost overly polished, but hey, some people are just fantastic at those things, and AI learns from actual people’s similar work too, right?

But now when I went to check, there are things that seem to be obviously AI generated in several pictures, mainly visible in the backgrounds. One weird Christmas tree, for example.

So I guess I’m he dumb one. It makes me sad because there are a couple of pictures I really like, which are not as over-the-top flashy as some of the others, and now I really question all of it.

@killer-laurent

I understand where you're coming from. It’s tough to navigate this situation if you're not familiar with how generative ai art currently looks. But trust me, it’s evolved far beyond basic filters that mimic oil paintings. Ai can now mimic a wide range of styles, often convincingly. That’s exactly why it's so important to listen to artists or those who can spot ai traits at a glance. Of course, it's also crucial to be wary of bad faith accusations, people wrongly accusing legitimate artists of using ai. That absolutely happens. However, the concerns surrounding Killer-Laurent aren't isolated incidents. I'll provide screenshots of the Christmas tree art you mentioned (and a link ) for full context:

As you pointed out, there are some oddities in the tree: objects blend together, melt into one another, or lack structure. This kind of "mistakes", especially along the tree's edges, is typical of ai struggling with complex designs. You see similar artifacts when ai tries to generate characters with intricate clothes/details. Ai ( at the moment ) just doesn’t handle that level of complexity well.

EDIT/UPDATE : It seems very likely that this tree was not ai generated but ran through an oil painting filter on photoshop. I will leave what I wrote intact for transparency but I will add a link here of a newer ask where this was brought up and clarified.

I should also state, before I get further into this part of my response: no, the issue isn’t simply that Killer-Laurent renders well. There are many highly skilled artists out there, and yes, AI does learn from real, talented creators. The real red flag is this: before he adopted his current super polished style, Killer-Laurent had a completely different way of drawing. There was no gradual evolution into this refined look. It just appeared. Suddenly, and without any signs of progression. Here are some examples of his earlier work:

These pieces look to have more hand drawn elements. For instance, Armand has a distinctive curl to his mouth that’s consistently depicted in his art from September and December 2023. Did he trace photographs at any point during these? Or use ai? It's very possible. At this point in time of his art, ai usage may be restricted to backgrounds ( the Christmas tree art was posted during this time period and the trees in the center image look suspicious)

[EDIT/NOTE: Those trees were also likely flittered through photoshop. Again, I am leaving my original statements for transparency but asking you to refer to the previous update for context).

I suspect, poses may be traced for some characters in much of his art during this period. There is a clear struggle with anatomy & fluidity in his art that have uncommon poses. Additionally, some character poses are suspect since he has shown earlier that year (June) to still be tracing photos of models. But that's a separate ( yet connected) issue. Ultimately, I feel these pieces have minimal AI ( if they have any at all.) But also, please note how he renders between these images. Very minimal, very subdued. And not very polished. To be clear: there is absolutely nothing wrong with being at this level in your artistic journey. I'm pointing this out to highlight what his art looks like when it seems he actually had a hand in creating it. And then, just days after that last set of earlier artworks, he posted this:

The first and last images depict huge leaps in both rendering and facial anatomy. Yet, the center piece is more consistent with his older work. This suggests a strange pattern: sudden improvement, a regression, then rapid progress again. Then by the time February arrives, he adopted three entirely new art styles:

Again, major improvements, all within just a few days. Then, four days after the third image, he posted this commission:

The female character’s face is more in line with his earlier style. I’m not sure if ai was used here, but I suspect he was forced to draw her face manually. That familiar curl of the lips, seen on previous art of Armand, appears here again, and it feels distinctly his style. However, the male character below (apologies, I forgot to screenshot the tags before the piece was deleted from his blog) looks different. The face structure is closer to the style found in his more AI suspect works, and it's clearly in a different style altogether. And then just three days later, he drew this:

Then, adopted several new art styles over the rest of the month:

Even his older artwork (below), where it does seem like he had a hand in creating it (maybe just coloring or touching things up), still raises eyebrows. Take this piece of Armand (the first image) as example. Besides the hair looking suspicious, looping into itself unnaturally, it also shares stylistic similarities with a specific style of ai art ( the three i included beside it) like there was an attempt to mimic the rendering style:

I'm not going to @ him in this one but I will link all previous posts about this. Alternatively you can look through my blog. First | Second | Third | Fourth | Fifth

At the moment, there are a few reasons that come to mind:

He provides content for the book-only side of the fandom. Not everyone in the book side of the fandom is happy with the changes made in AMC's recent adaptation. Many are still deeply attached to the original descriptions from the books. But as the show gains popularity, content tailored specifically for book fans has become increasingly rare, if not completely drowned out in the tags. That scarcity creates a kind of hunger. And when you're starved for content, you'll take what you can get. I think that’s why some people continue to support him because at least he’s giving them something.

It looks professional. The rendering quality gives off the impression of someone with years of experience. When you combine that level of polish with the lack of similar content, it’s not surprising that it satisfies a certain niche. It looks good and it fills a gap.

They believe the timelapses. There’s still this notion floating around that timelapses can’t be faked --- that they’re undeniable proof of a manual process. Add to that the fact he posts WIPs, and people start thinking, “Well, clearly he’s showing his work, so he can’t be using AI.” But it’s not that simple.

A false sense of loyalty. Some people either are his friends or think they are. Maybe they commissioned him once, or he drew for their favorite ship/character. And because he’s “one of theirs,” they feel obligated to defend him against criticism. The logic seems to be: if you dislike what he’s doing, you must be jealous. And honestly? If there are jealous artists out there, I don’t even blame them at this point.

@killer-laurent

I imagine this is true for many in the community who didn't realize his art was ai generated. It's one of the main reasons I always use fandom tags, to ensure my posts reach more people, especially those who trust the support he's received. It seems like only artists, who have more at stake in speaking out, acknowledge that his use of AI isn't just questionable but blatantly obvious. Some of the more popular blogs in this fandom space are run by writers, which I understand might make it harder for them to spot AI art at a glance. But I hope they realize that by continuing to support AI - especially in such a niche fandom - they may unintentionally harm the genuine artists involved.

For example:

1. You are actively making artists compete against ai for commissions when you promote them in the same space. 2. This could increase the anxiety that artists already feel about ai scraping their work, and in the long run, you risk losing valuable creators from this fandom. 3. You might also alienate artist friends who feel betrayed by your defense or lack of acknowledgment of his ai use. 4. Beginner artists, unaware that his work has ai because they trust you, may see his work being promoted and think it's achievable to work at that level of output and schedule. This can lead to burnout, and the physical toll it takes on their health should not be overlooked.

I never said using stock images was wrong. I verified the information, provided evidence that it’s possible to replicate similar results (which was the entire point, to remove suspicion of AI), and I corrected myself because I made a mistake. I made sure to include the tags because I wanted everyone to see the update.

After all, If I had publicly claimed the art was ai generated and later found out it wasn’t, wouldn’t it be reasonable, even necessary, to correct that publicly, just as I had shared the misinformation publicly? It would’ve been disingenuous to quietly update the post without tagging it, leaving others to manually find the correction.

Yes, artists can and do change styles frequently. And when they do, you can also tell it's still them drawing it. But you’re missing the nuance here, the kind of style shift I’m referring to isn’t just exploration or growth. It’s sudden leaps in skill and technique that happened within days. That’s not common.

I get that people are tired of seeing the same discourse, trust me, so am I. But what’s more exhausting is watching a fandom pretend nothing is wrong while continuing to reblog and support someone who’s clearly using AI. It discourages other artists from sharing their work in the same space, not because of how annoying I am (because people weren't listening long before I came into the picture), but because they see the fandom openly backing someone who’s cheating. It sends a message that in a fandom that’s hungry for content, people can get away with faking it, and no one will care.

Here is a link to the post this ask is referencing.

Hey, thanks for pointing this out! I did a bit of digging, found the tree (pretty easily, like you said), and made a comparison. And yeah, I think you’re absolutely right. It looks like he applied a filter to it.

That said, I still have a few suspicions since there are parts that don’t quite line up. But I’m willing to let that go, since I don’t know the exact process he used with the filter.

Here’s some context with screenshots:

First is Killer-Laurent’s tree. I covered the characters so we can focus just on the tree. Second is the original PNG of the Christmas tree.

I tried to replicate the effect Killer Laurent used by running the original tree image through Photoshop. I used the Oil Paint filter in PS CS6, followed by Poster Edges from the Filter Gallery to add the outline effect.

Here’s the result:

Even after a few rounds of filters, I couldn’t get the same level of distortion as in Killer Laurent’s version. The edges especially, and the top part of the tree, where the needles look more like clumps of hair, are what stand out to me as particularly odd. That’s the part I still find a little suspicious.

It’s still within the realm of possibility that he ran the image through ai and prompted it to look like an oil painting. But based on how close my filtered version looks, I’m inclined to believe he just used filters and mirrored the left side of the tree onto the right, especially since it’s clear he didn’t just flip the whole image.

Here’s what it looks like with the sides mirrored. I think it comes pretty close, except for those weird distortions in the ribbons and pine needles. But the color blending and object merging are pretty similar.