Yet another topic about AI… 🫣 It’s been about a year since I started using Glaze and Nightshade on my artwork images before posting them online, and I was assuming that this would keep me somewhat safe from AI unethical appropriation and scraping. For you guys who aren’t familiar: Glaze and Nightshade are softwares currently developed by researchers at the University of Chicago in order to protect copyright and safeguard artists from invasive AI tools. Glaze, specifically, promises to impair style mimicry, and Nightshade promises to “deceive” AI models so that these models won’t be able to “read” the protected images. Then, I thought I’d run a few quick tests. TEST 1 - I applied Glaze (at its default level) to image number 1 attached here (sketch of a lady with cats), uploaded it to ChatGPT and asked it to generate an image of the Mona Lisa in the same style. Theoretically, it shouldn’t be able to do it, since Glaze is meant to prevent style mimicry. But take a look at image number 2 and tell me: do you think Glaze’s protection actually worked in this case? TEST 2 - I apele Nightshade (at its default level) to image number 3 (sketch of a guy in glasses with a backpack, running, with a thought bubble where’s writtn “Where’s the bus?”), the I uploaded it to ChatGPT as asked it to describe in words what was that image. It shouldn’t be able to do it as well, since Nightshade’s promise was to “poison” the image so that it would be “unreadable” to the AI tool. Yet, image 4 is what ChatGPT gave to me as an answer (since the text is in Portuguese, here’s a general translation: “The image is a cartoon-style drawing made with loose pen strokes. It shows a seemingly young person wearing large glasses running with a backpack on his back. The character appears to be in a hurry, perhaps trying to catch public transport. There is a speech bubble with the phrase in Portuguese: "Where's the bus?" This indicates that he is looking for or waiting for a bus that has not yet arrived. The style of the drawing is expressive and humorous, suggesting an everyday scene with a comic twist. There is also a small signature in the lower right corner, with the letter "B". Would you like me to apply any modifications or create something based on this image?” So I wondered if Nightshade worked only on my signature (which is “LI”, not “B”), since that’s the only part which the AI wasn’t accurate about… Would a higher level of Glaze/Nightshade protection perhaps work more efficiently? (I hope I'll be able to test these higher levels in the future, although I couldn't up to now because applying the standard level of Glaze and Nightshade to a single image costs me several hours, literally… I wonder if my PC is too slow.) I truly don't mean my words to be a criticism to the work of the brave team of researchers at the University of Chicago who, I believe, are doing the best they can with Glaze and Nightshade. I honestly and deeply appreciate their endeavour. It's just that, after all this, I have doubts and wonderings: - Is any artwork ever really “safe” online nowadays? - As artists, is it efficient or even worth the effort to try to “safeguard” ourselves against AI through such strategies? Assuming that the problem is not the tool, but how people use it, is it too naive to trust that it’s possible to raise ethical awareness for people not to use AI maliciously in the first place, or at least to trust that those who do will face effective legal consequences? - Even if it was possible to completely avoid AI, if AI is where the future of overall digital information is headed, wouldn’t the attempt to “protect” ourselves against it backfire? In other words, by trying to keep our artworks out of AI databases (if this is even possible by now), would we be excluding ourselves from the future of digital memory/history? Among all this, I even believe that art and copyright are just a fraction of the problem - deepfakes, data manipulation, social control, political surveillance, these are much more complex issues also implicated in the "bad side of AI". Still, it’s somehow all connected, isn’t it? I know this is a polemic and even difficult matter for many of us. And I know that answers or solutions are complicated. Anyway, I’d love to hear other thoughts about this. Stay well, y'all! 🤍

A couple days ago, I got an e-mail from an artist whose work I've admired for long and from whom I hadn’t heard in a while. He used to call himself an "art hacker" and seemed to be what many of us would call "succesful" - he had good jobs in the entertainment industry, got well paid for them and was able to create amazing illustrations with innovative concepts, solid construction and beautiful rendering. Then, around 2020, at what he said was the peak of his career, he quit. No, he didn't die - in fact, he said he was "alive and well". But he did quit his art jobs, "killed" his artistic persona, completely left the art industry and openly donated 90GB of his stunning digital artwork to public domain. He said it was a deliberate decision of "creative destruction" in face of realizing that, basically, mass-production entertainment was a flawed institution in our society. By quitting his career as an industry-standard concept artist, he said he could free his creative energy from restraints and direct it to fulfill other objectives. As generative AI came around and started becoming more popular in the latest years, he wrote again. He said “AI is killing art” and seemed very concerned not only about the future of art jobs, but also about the future of humanity and modern society overall. And below is the most recent message I got from him: — “By the end of 2025, most traditional artist jobs will be gone, replaced by a handful of AI-augmented art directors. Right now, around 5 out of 6 concept art jobs are being eliminated, and it's even more brutal for illustrators. This isn't speculation: it's happening right now, in real-time, across studios worldwide. At this point, dogmatic thinking is our worst enemy. If we want to survive the AI tsunami of 2025, we need to prepare for a brutal cyberpunk reality that isn’t waiting for permission to arrive. This isn't sci-fi or catastrophism. This is a clear-eyed recognition of the exponential impact AI will have on society, hitting a hockey stick inflection point around April-May this year. By July, February will already feel like a decade ago. This also means that we have a narrow window to adapt, to evolve, and to build something new. Let me make five predictions for the end of 2025 to nail this out: 1 Every major film company will have its first 100% AI-generated blockbuster in production or on screen. 2 Next-gen smartphones will run GPT-4o-level reasoning AI locally. 3 The first full AI game engine will generate infinite, custom-made worlds tailored to individual profiles and desires. 4 Unique art objects will reach industrial scale: entire production chains will mass-produce one-of-a-kind pieces. Uniqueness will be the new mass market. 5 Synthetic AI-generated data will exceed the sum total of all epistemic data (true knowledge) created by humanity throughout recorded history. We will be drowning in a sea of artificial ‘truths’. For us artists, this means a stark choice: adapt to real-world craftsmanship or high-level creative thinking roles, because mid-level art skills will be replaced by cheaper, AI-augmented computing power. But this is not the end. This is just another challenge to tackle. Many will say we need legal solutions. They're not wrong, but they're missing the bigger picture: Do you think China, Pakistan, or North Korea will suddenly play nice with Western copyright laws? Will a "legal" dataset somehow magically protect our jobs? And most crucially, what happens when AI becomes just another tool of control? Here's the thing - boycotting AI feels right, I get it. But it sounds like punks refusing to learn power chords because guitars are electrified by corporations. The systemic shift at stake doesn't care if we stay "pure", it will only change if we hack it. Now, the empowerment part: artists have always been hackers of narratives. This is what we do best: we break into the symbolic fabric of the world, weaving meaning from signs, emotions, and ideas. We've always taken tools never meant for art and turned them into instruments of creativity. We've always found ways to carve out meaning in systems designed to erase it. This isn't just about survival. This is about hacking the future itself. We, artists, are the pirates of the collective imaginary. It’s time to set sail and raise the black flag. I don't come with a ready-made solution. I don't come with a FOR or AGAINST. That would be like being against the wood axe because it can crush skulls. I come with a battle cry: let’s flood the internet with debate, creative thinking, and unconventional wisdom. Let’s dream impossible futures. Let’s build stories of resilience - where humanity remains free from the technological guardianship of AI or synthetic superintelligence. Let’s hack the very fabric of what is deemed ‘possible’. And let’s do it together. It is time to fight back. Let us be the HumaNet. Let’s show tech enthusiasts, engineers, and investors that we are not just assets, but the neurons of the most powerful superintelligence ever created: the artist community.” — Now I’m a fan of creative debates, so I figure bringing this discussion here could help put our minds to think together. So what do you guys think? Do these predictions make sense to you? Do you think “AI is killing art”? Do you feel empowered enough as part of an art community to “hack the future” or so? And how do you feel about this whole scenario we’re currently living and witnessing? Stay well! <3 And let’s share ideas!