First let me say that I hate AI slop and the concept of using AI overall, even at work, because the way that companies dictate its use just takes away the soul of the work, the things that I liked about the work. The general economic sentiment is that output must be maximised and this leads to different problems in most cases. Nowadays, however, some people and companies alike started to realise that maybe the current implementation model is just not the right way. Nevertheless, despite the overwhelming AI slop that drowns the internet, there are some AI made contents, such as animation series or art, which portray that there can be genuine ideas behind it like the overall setting and the characters in the said series. And thus this proves that when used right with true human input, it can become an AI project that people enjoy. I really think that these things could work. Be it for example, a short form of animated narrative that is actually captivating. And thus people would watch and carve more of it like in the case of any other series or cartoons hosted on streaming platforms or on the internet.
I do not want to fall into the common trap of absolute binary thinking (i.e., “AI is pure evil” vs. “AI is the future of everything”). Slop existed long before AI, from cheap paperbacks typed on typewriters (mind you when I go into the book store nowadays, I see a lot of slop too…) to mindless content streams on social media. In this manner technology should not and cannot be blamed. Instead, I think that the blame should be placed squarely where it belongs: on human intent, effort, and execution. Consequently, tools merely amplify the user’s intent. A low-effort user produces slop regardless of the medium; a skilled craftsman produces art. And I put these words here with my own scepticism and frustration with corporate-mandated AI usage in mind. So by no means I try to defend AI use, but I think that there is a difference and effort should always be recognised and commended.
Generally, I am critical of AI use, but at the same time I use AI too. The difference between the overall majority of AI users and the talented few who can create original content and wonders with it is the approach. AI, like any other tool, is just a tool. It can be used in both bad and good ways. Unfortunately, the internet is full of slop, but it was no different before AI either. For a while we had TikTok slop, Instagram, YouTube or even Photoshop slop too. When it comes to real human creativity AI can be a tool to realise it. However, many people use it wrong. This series can prove that good things can be made with it, when the right person uses the tool, same as it is with TikTok, Photoshop, or even with typewriter… you can write slop with a typewriter too and you can also create something like the Lord of the Rings or The Witcher with it too… The thing is that the internet is overwhelmed with slop and thus we perceive the use of AI as being unethical or bad mostly, because of how the majority of people use it. The main point here is that we need to have a critical approach to AI and its output, we cannot just accept what it spits out without a grain of salt.
Essentially, AI slop is the result of passive execution—using tools to generate maximum volume with zero effort or quality control. The opposite—which I would like to call curated AI or “craft” with AI—requires high human input and leverage. Namely, intentional prompting, domain expertise, rigorous editing, and creative direction are must haves that ensure AI’s role as a tool and not the “creator”. And I genuinely think that this is the mistake that most people make.
We need to have a critical and creative approach to AI. It is not enough to outsource everything to it. If there is human input and real imagination behind its usage, then AI can act as a support or an assistant who makes your job easier or complements the exact thing that you lack, for example the skill to animate. However, when you write a prompt in a very complex way and put the effort into building up the creative workflow, the human will be behind the end-product and it won’t feel as an AI slop any-more.
My criticism and in general my critical approach also takes into consideration the ethical parts of this discourse and the overall AI usage. Mostly, it does not matter how meticulously you curate the output or how complex your prompt is, the underlying models were trained on millions of artists’ and writers’ copyrighted works without consent or compensation. Using AI—even as a thoughtful craftsman—is still built on exploited labour. Which is a valid critique of how current commercial models were built, and copyright regulations urgently need to catch up. However, the ethical issue of dataset sourcing is separate from whether generative tools can function as a creative medium. You can also build your own generative models that you train, for instance on your own work on your own local computer. Same as you train yourself by looking at art made by others and by gathering inspiration from books or films or whatever medium you prefer. If a model is trained on ethical grounds, whether it is based on your work or licensed and open-source data, the core premise remains: human direction and craftsmanship are what determine whether the final output has soul or is just slop. Human culture in general is the circulation and reframing of ideas about which Michel Foucault famously wrote his essay “What is an Author?” (Qu’est-ce qu’un auteur?) in 1969, way before AI. He wrote it two years after Roland Barthes proclaimed “The Death of the Author” (1967) as a response.
In summary, Barthes argued that the author’s biography, psychology, and intended meaning should no longer control or limit the interpretation of a text. By “killing” the author, Barthes aimed to liberate the reader and open up endless possibilities of meaning within the language of the text itself. In addition, Foucault shifted the question from what the author means to how and why we use the category of the “author” in society. Instead of just celebrating the author’s disappearance, Foucault in his essay investigated the institutional and cultural mechanisms behind how we treat texts. Namely, attaching a name to a text functions as a tool for society to regulate, classify, and limit discourse. Not all texts require an author (e.g., shopping lists, everyday notes, or legal contracts usually do not), but discourses like literature, philosophy, and science use the author’s name to give weight, ownership, and coherence to a body of work. However, most of these texts are inspired by previous texts and ideas from others. This is how culture builds on itself. Our Western culture places too much emphasis on individualization. In this sense, the author represents a specific moment of individualization in the history of knowledge—a way of attributing responsibility and coherence to ideas. Foucault famously suggested that the figure of the author might eventually dissolve, leaving behind only an anonymous trace or murmur—asking not who is speaking, but simply “What matter who’s speaking?”—which is put into a totally new light under the current emergence of AI content.
Naturally, using AI to bridge skill gaps, like animating a story, can be criticised vividly and justly so. Learning to draw, animate, or compose is where artistic expression lives. By skipping the physical struggle of mastering a medium, people lose the happy accidents, personal style, and genuine humanity that make art meaningful. However, technique and vision are not the same thing. Throughout history, technology has removed technical barriers so more people could focus on vision. Directors no longer draw every frame of a movie; sculptors historically used assistants to rough out stone blocks. AI allows a storyteller with deep vision—who might lack 10,000 hours of manual rendering skill—to bring a world to life. The ‘struggle’ simply shifts from manual execution to vision, story-building, and meticulous editing.
There is also the problem of describing a vision through prompts and tweaking parameters. These are directorial, not thus cannot be categorised as craft. In this instance the user is a commissioner or an editor, but not the creator. Therefore, calling it ‘craftsmanship’ dilutes what craftsmanship actually means. As a result it is important to differentiate between the models and forms of AI usage. So, that is exactly why terms like ‘Curated AI’ or concepts, such as ‘AI-assisted creative direction’ fit so well with my argumentation. Curation and editing are crafts in their own right. For instance, film editors do not shoot the footage, but their cuts define the narrative strength of a movie (not to mention that even cutting is done nowadays by a dedicated person and not by the director!). Recognizing AI output as raw material that requires heavy human refinement does not diminish traditional craft—it defines a new layer of directorial skills.
By approaching AI with craft or craftsmanship, I mean that we should use AI with intention, meticulous fine-tuning, and actual artistic skills—rather than just throwing a vague prompt at a generator. It highlights the process, effort, and deliberate human control behind the result. As a result, AI becomes an assistant and thus a useful tool that is not replacing real human effort, but amplifying human capabilities. This approach consequently positions technology as a collaborative partner—one that expands our creative boundaries and tackles problems beyond our individual reach. So this thought process leads me to the phrase “curated AI” that highlights the critical role of human oversight, selection, and refinement. Great AI work rarely comes straight out of the box; it requires human taste to supply the foundations, to build the creative background and then to filter, edit, and polish the raw material that AI provides based on these grounds.
Of course, by being an idealist as always, my idea of ‘AI Craft’ sounds noble… too noble to be true, because the market does not reward craftsmanship—it rewards volume and cost-cutting. Companies won’t pay for slow, curated AI projects when they can flood channels with 1,000 auto-generated pieces for a fraction of the price. In this case, slop is not simply an accident; it is the economic goal. This is exactly at the core of the current problems that we face and why the internet is drowning in slop content. Nevertheless, it was, as I have emphasised before, not so different before AI either. Short-term corporate incentives always favour cheap volume—we saw the same race to the bottom with clickbait articles and cheap content farms. But audience fatigue is real. As the market becomes over-saturated with generic slop, high-quality, human-curated content actually becomes more rare and valuable. People eventually tune out noise and seek out depth.
In summary, good content can only be made with AI if there is real value and human effort behind it. So, when AI is treated in its place, as a tool. Therefore, using AI with the mindset of a craftsman who utilises a tool, describes work where AI is used to handle heavy lifting, while human expertise guarantees accuracy, depth, and genuine insight. And I think that this can be the key to real and genuine content that was made with the assistance of AI.