# AI Slop vs. Canva Slop: Why AI Design Tools Make Everything Look the Same [00:00:00] Okay, so I want you to picture something, and I bet you can picture it right away, 'cause you've probably seen it, like, 50 times this month or more without even registering it. It can be either a flyer or a poster or, you know, it doesn't really matter what it's for. [00:00:22] It could be a webinar. It can be a local event. It can be an ad, Instagram ad for a local business. It can be someone's course launch. And, you know, you have that bold headline with a lot of elements floating around in the air, some vague abstract shape floating somewhere it doesn't need to be, and then text that is super busy and not quite aligned the way that designer would necessarily align it. [00:00:59] And you know the one. You know what I'm talking about because you've scrolled past a dozen of them just this week alone. And I started noticing it everywhere, on YouTube thumbnails, on local flyers for a coffee shop, LinkedIn event graphics. It's like a specific visual dialect that has appeared out of nowhere, and now it's just everywhere. [00:01:22] And I had this weird sense of deja vu about it because I watched this exact pattern happen before, and it just wasn't necessarily AI slop back then. It started when Canva became popular [00:01:58] Hi there, creative peeps. So today, I wanna talk about something that's more observational than necessarily instructional. So there're not gonna be any tool tutorials today or step-by-step guides. I wanted to name a pattern that I think a lot of you are noticing too, and I wanna think through aloud what it means 'cause I don't think it's necessarily, okay, this is super bad, or okay, it's fine. [00:02:26] J- you know, just get over it Let's go back for a second. When Canva really took off, properly took off, not just when they launched, I mean, you know, after that early adopter phase, something specific happened. The friction of making something that looked designed basically disappeared. Because before that, if you wanted a halfway decent flyer, you either had to know design software or you knew someone who did, 'kay? [00:02:57] So there was a gate. It wasn't a huge gate, but there was one, right? And it meant that most of what got made by non-designers looked like what it was, a little rough, a little homemade, but at least it was trying to be its own thing. But then Canva removed that gate, and what filled the space was not suddenly a world of better design. [00:03:21] There was a flood of the same templates, you know, slightly recolored maybe, over and over until you could spot a Canva template across a room, right? And we all developed that instinct. You know, you'd see a particular gradient and particular font pairing or a layout with a circle behind the headshot or whatnot, and you would just know. [00:03:50] And a lot of the templates were pretty good, you know? They were designed by graphic designers at Figma, or I'm sorry, at Canva. But the problem was the scale of usage. So there's one good template, and it's multiplied across a million users with zero customization instinct, and the template stops looking like good design and just starts looking like a uniform, right? [00:04:19] There's something interesting I found out while I was researching for this episode. So this pattern, it's not necessarily related to Canva, and in fact, it's got a name, and it goes back way further than I expected, and it's called the ransom note effect. That's a typography term, and it describes layouts with way too many fonts crammed together, so there's really no hierarchy. [00:04:49] There's clashing weights, and it looks like a message cut and pasted together from letters in different magazines, and it's named that because that's literally what it looks like, the classic cut-out letters kidnapper notebook. And that term comes from the 1980s from desktop publishing, in fact. So when the Macintosh, the laser writer printer, and the program called PageMaker came together in 1985, it was a massive shift, right? [00:05:18] Because for the first time, anyone with a computer could lay out A proper document, right? They could choose fonts and then place images, basically control a whole page without going through a professional print shop. It was that Canva moment, but decades earlier. So same exact promise, tools previously locked behind professional training suddenly become available to anyone. [00:05:44] And the result, almost immediately, was a flood of chaotic-looking flyers and newsletters and church bulletins because people had access to a dozen fonts for the first time in their lives and used every single one of them on the same page, you know, just because they could. And there's actually a documented quote about this from the software industry at the time. [00:06:08] The drawback of putting professional publishing tools in the hands of people with no design training was that, quote, "All too often users mixed all sorts of things together just because they could." And get this, the exact same criticism got leveled again, word for word almost, about a decade later at early web designers. [00:06:30] So when HTML and basic web tools made it possible for anyone to build a website, the early internet went through its own ransom note phase. You know, the blinking text, 17 fonts, backgrounds that fought with everything on top of them. You know, same pattern, but just, you know, in a different decade and with a different tool. [00:06:50] So this is now three times, right? Publishing, desktop publishing in the '80s, then early web design in the '90s, let's say Canva in the 2010s, and now ChatGPT Image 2.0 doing it again with AI visuals in 2026. And there's actually a fourth one sitting right in the middle that I forgot to mention in the timeline that I think is most useful comparison of all because it was the opposite problem, and it was with WordPress. [00:07:21] So WordPress didn't really cause a ransom note moment because most WordPress themes are, are well designed individually, right? Because they're made by skilled people. And here we had a different problem. So it's the sameness of it. Not really chaos as in other examples. When you scroll through small business websites in pretty much any industry, and you will see the same hero section, and then the same three-column service grid, um, then the same testimonial carousel, same stock footage or photos of someone shaking hands over and over. [00:08:01] You know, just different colors basically, uh, swapped in. And people in web design talk about this constantly. There are entire articles just titled some version of, "Why do all WordPress themes look the same?" And that's, as a matter of fact, the more accurate preview of where AI image generation is heading, I think more than the ransom note chaos. [00:08:26] If you remember, early ChatGPT images had that chaotic, slightly broken look. It was super bad text, right? Weird blobs and hands looking super weird as well, and compositions that didn't really make any sense, and that was the messy phase. But Images 2.0 is pretty smooth and competent. I use it all the time now And it's super smooth if you wanna visualize something, right? [00:08:56] Super professional-looking on the surface, which means we're not heading back into ransom note chaos, but towards the WordPress kind of sameness, right? So it's less obviously bad looking, but more so identical. And there's a third failure mode sitting between those two, and I honestly think that's the most common one of all, and that one I will mention separately. [00:09:22] And it's someone taking a really well-made template, whether it's good Canva layout or a really sleek WordPress theme, whatever it is, and then wrecking it themselves through formatting. You know exactly what I mean if you've ever scrolled past it. Someone starts with a well-designed template. There's clean spacing, intentional hierarchy, font pairing that works. [00:09:50] And then they go in and swap the headline font for something that doesn't really match, and then they center text that was supposed to be left-aligned because centered feels more important for them. And then they cram in three extra paragraphs because they had more to squeeze in there, and then the spacing that made the whole thing breathe just collapses, right? [00:10:14] And then they change the brand color to something slightly off because it's closer to their brand color, except now it clashes with everything else in the template that was built around the original palette. So it comes down to that the person using it didn't have the design knowledge to know what changes would break it and which ones wouldn't. [00:10:39] Because they don't know why the spacing was doing what it was doing, so they don't know that they're undoing it, right? They just know that they want their logo bigger and they want to fit one more bullet point in or whatever, and then they make that change without any sense of what it costs the rest of the layout, right? [00:11:02] And I think this is the most relevant failure mode for AI image tools specifically because of how these tools work. Images 2.0 generates something coherent, but then people regenerate it five times trying to get that one specific tweak right, or they take the output into Canva afterwards to add their own text on top because the AI tool wasn't quite... [00:11:26] didn't quite turn out right. And that's exactly the moment where someone with no design training starts making choices the original generator never accounted for. So the tool handed them something that was pretty decent, but they just don't have the eye to leave it alone or to know which part to touch, okay? [00:11:48] So that's what happens every single time when the cost of access to a creative tool drops faster than the cost of developing taste, right? Whether that shows up as chaos, like we mentioned with the ransom note effect, or as sameness, like with WordPress, or as good work ruined by someone who didn't know what not to touch. [00:12:09] Because acquiring taste is slow. It comes from looking at thousands and thousands of examples, making mistakes, and having someone tell you why something doesn't work. That's the whole learning process. But access can drop to zero in a single product launch, right? And whenever that gap opens up, you know, fast access and slow taste mixed together, you get a flood of mediocre output because everyone's filling the same gap in their own judgment with the same defaults or the same well-meaning bad edits Okay, so now we've got ChatGPT's Image 2.0, which dropped this April, and it's a big jump, definitely a big jump from where AI image generation was even a year ago, or even six months ago. [00:12:57] So essentially how it works, it reasons through a prompt before generating. It can search the web mid-process, and it renders readable text now, which used to be the thing that gave AI images away instantly before. You know, do you remember when every AI poster had the gibberish text that looked like letters, but wasn't? [00:13:18] And now that is mostly fixed, and it can definitely produce, I think, up to eight or even more coherent matching images from just one prompt, which means, like, an entire matching set. In- Instagram, a carousel, story, banner, thumbnail, all generated together, all looking like they belong in the same campaign. [00:13:41] And it's free or close to free for a huge number of people. That's the part that mirrors Canva most closely. The thing that used to cost a lot of money or hiring skilled designers is now a few seconds away and a sentence away, right? Here's where I relate it to Canva because th- you know, th- frictionless tools in the hands of people with no design training produce something that looks like it's well designed on the surface, but then it falls apart the second you take a deep look at it. [00:14:17] Those are all compositions that are super busy because the tool is good at filling space. And here's where I think it's a little more concerning. You know, Canva templates were built by human designers, and then they were copied infinitely. The starting point was good, even if the multiplication wasn't, okay? [00:14:38] But AI image generation doesn't have that same anchor, so it's pattern matching across millions and millions of training images, which means what comes out is an average. It's a blend. So it's drawing from everything at once instead of one specific internal point of view, and that's the real difference. [00:14:58] Canva slop was one good idea overused, but AI slop is no specific idea, but it's executed smoothly. Okay, in all fairness, I think for most of us who care about design, we can easily slide into being super precious about this topic. But for most people making these posters, flyers, thumbnails, whatnot, the bar was, this exists, it communicates the basic information I need to put out, and it costs me next to $0 and three minutes instead of $500 and three days. [00:15:35] Okay, let's say you are a local events organizer, and you are throwing a fundraiser. Well, basically, you just really need something legible that says when and where, you know? So AI image tools are, in fact, solving a real problem for a huge number of people who were never gonna hire a designer in the first place. [00:15:57] So that's not a problem. But I think what's worth mentioning is what happens at the edges, okay? So small businesses who used to hire freelance designers for their event flyer, because to them $200 felt worth it, and now don't, because four minutes and free feels like the same outcome for them, right? So if you are a working designer, or honestly, if you're just someone who looks at a lot of visual content, you start to develop the same instinct we all developed for Canva. [00:16:32] You can spot it. You can spot the too clean gradient and composition that's busy in a way that has no hierarchy behind it. Text that's legible, but it's sitting in a spot that fights the rest of the layout instead of working with it. And I think that tell is valuable as a way to understand what real design is doing that the tool still is not. [00:16:56] Because good design is a series of decisions, each one ruling things out. So AI image generation doesn't rule things out the same way, but it's actually optimizing for looking coherent and complete. Not every choice earns its place on the page, right? I don't know if that makes sense, if I, I explained that clearly. [00:17:20] So essentially, the gap between looking complete and looking intentional is, I think, where now hiring a real designer lives in this specific moment in time. So, okay, access to tools is not a problem anymore. Everyone has it. But it's really about when you have someone who can look at a blank page and make 20 small decisions that add up to something with a, a unique point of view instead of an average of every poster that's ever existed. [00:17:51] So if you're a designer that is aware of this, to be honest, I think the move is, as in most instances, somewhere in the middle. So it's not like, okay, full panic mode, this is total BS, or okay, let's just close our eyes and pretend nothing changed either, right? You need to get specific about what you are selling, okay? [00:18:16] When you think about it, it's the 20 ruled out decisions. And when you start naming that explicitly to clients, your whole process, I think that will give them a better understanding to know what's the difference they're paying for. So use these tools that we have at our disposal for the first draft, but not for the final output. [00:18:39] The same way we use AI for writing. So you are generating a rough draft, but then the actual decisions are layered on top of it. And it's the same with image generation. And as a matter of fact, it's an even faster path to good work. And if you're someone who is not a designer and you're using these tools for your own flyers, thumbnails for YouTube or whatnot, I mean, that's fine. [00:19:05] That's what they're there for. Just know that when, you know, stakes go up and you're either growing and you're investing in your brand or you have a campaign that needs to convert, anything where the visual is actually doing the strategic work, that's the moment where the average is not good enough, okay? [00:19:26] Oh, and one more thing, specifically for that third failure mode. So if a tool hands you something that already looks pretty clean and finished, resist the instinct to immediately go and fix one thing, okay? Because sometimes the best move is to leave the well enough, quote unquote, alone, especially if you cannot articulate exactly why the layout is working in the first place. [00:19:56] If you absolutely don't know why something looks balanced, that's usually the moment to stop touching it and not the moment to go in and make it yours. Okay, so to wrap this up, I think we're going through the same story we've already lived through many times, but just a lot faster and with a wider gap between the floor and the ceiling. [00:20:20] And I wanna know, what's your take on these, on these AI posters, thumbnails, flyers? Do you think it's just a temporary thing and then everybody's gonna get sick of it and move on to something else? Or do you think it's gonna impose a bigger problem? All right, that's it for today. Thank you so much for listening to another episode of AI For Creative Professionals. [00:20:44] If you haven't already, I would really appreciate it if you leave a rating or a review on Apple and Spotify. It would really mean the world to me, and it would help more creative professionals like you find this show. Talk to you next week.