# AI is not a tool - AI & SOCIETY **By:** Nataliia Laba **Published:** 2025-12-29T00:00:00Z **Source:** [Springer London](https://link.springer.com/article/10.1007/s00146-025-02784-y) --- Dear Editor, Promotional rhetoric of tech companies frames AI as a revolutionary technology poised to unlock or augment human potential. This framing is particularly evident in the domain of cultural production, where visual generative AI—technology that manufactures synthetic images—is marketed as the definitive solution to _your_ creative needs. I argue that there is a tripartite structure to these slop slogans: 1. 1. Creativity is defined as a territory to be conquered. 2. 2. AI is presented as a tool for this conquest. 3. 3. Present-day concerns are dismissed as obstacles to an inevitable, AI-driven future. ## 1 The territory: creativity as a problem According to AI companies, creativity is not an innate human endeavor, but a problem, an inefficiency waiting for a technological intervention. With Midjourney, for example, one will explore “new mediums of thought” that expand “the imaginative powers of the human species” ([https://midjourney.com](https://midjourney.com)); and for Stability AI, the future of creativity is achieved by “people like you, using **tools** like ours” ([https://stability.ai](https://stability.ai)). In this framing, “the imaginative power of the human species” is implicitly presented as a resource to be optimized, where the “future of creativity” is something to be delivered by an AI tool. Such rhetoric conveniently abstracts creativity from a complex, embodied human experience and represents it as a computational problem to be solved. We have seen this techno-solutionist playbook in countless technology cycles: Facebook promised to streamline social connection; Adobe emerged to optimize the drudgery of manual production; NFTs claimed to solve digital ownership; the Humane AI pin sought to liberate us from screens; Apple Intelligence offers to unburden our minds from information overload. The list goes on. What unifies these and so many other tech companies is an unwavering conviction that complex social and personal issues can be neatly solved through technological intervention. Ontologically, we are shown that (a) given societal configurations are undesirable, and thus (b) we must embrace the radical alternatives brought about by AI. The core problem presented by AI companies selling access to image generation systems goes something like this: ‘I have an idea in my head, but I lack the skill, time, resources, or commitment to make it real.’ The deeply personal nature of creative work is either reduced to a set of needs that can be met with an AI solution or augmented by AI not as a distant, disembodied technology but as a creative partner, collaborator, companion—a computational friend. Creativity is a territory to be entered, mapped, traversed, and ultimately, managed by the technology on offer. By defining this territory, AI companies create the market for their solutions. ## 2 The tool: AI as the solution Once creativity is established as a problem-space, visual generative AI becomes an indispensable _tool_. This “tool” is everywhere. It is infiltrating LinkedIn through AI headshot generators, promising to replace expensive professional photography. It lingers on platforms like Buffer and Hootsuite, offering AI-powered caption generation to craft perfect posts. It is on design platforms like Canva, waiting for us to generate assets to elevate imperfect human communication. The tool is quasi-magical—enigmatic and unpredictable—with images (or so-called “AI art”) ritualistically emerging from prompting séances. What the tech industry wants us to think about is creative possibilities, from generating personalized financial auras (e.g., Intuit’s Automagical campaign) to copying the likeness of other people (e.g., omnireference in Midjourney’s version 7) for ‘character consistency.’ Such generation processes co-opt the language of mysticism, a blend of automation and supernatural wonder. Where artist Paul Klee talked about art as making the invisible landscapes of the mind visible, visual generative AI works in reverse—it packages renderings of invisible algorithms as art. Yet what remains invisible is precisely what we should be interrogating: how, when, and which parts of visual generative AI are involved in these practices. These are not neutral technical questions. AI offers regurgitated epistemological positions from its training datasets, model architecture, and system affordances, all of which embed assumptions about how social realities should be represented. By accepting that AI “sees” for us, we are presented with the world whose epistemological positions have been predefined by corporate elites in pursuit of market dominance. Sure, the user is free to create so-called art with the so-called tool, but only within the confines of a pre-digested world. > A way of dramatizing the point more polemically is to say that visual generative AI is to art what paint-by-numbers was to painting. Rather unfortunately, the tool metaphor is a powerful rhetorical strategy repeated and amplified at enough volume and scale. It begins to feel like an objective truth, even within institutions that should be its most critical examiners. This complicity manifests in several ways, from the language used in academic grant proposals that frame AI as a novel tool for solving complex social problems, to the rise of AI ethics as a field that, while important, may legitimize and naturalize industry discourses by focusing on mitigating AI’s harms rather than questioning its fundamental premises. Adopting this language, AI critiques inadvertently lend intellectual credibility to the industry’s preferred narrative, undermining their own intent to expose the very power structures they set out to challenge. ## 3 The future: temporal deferrals and the yet-to-come Superimpositions of human inefficiencies as a problem and AI as a tool are aimed at eradicating scepticisms and adoption barriers. The complexities of, for example, mastering the technique of collage—from sourcing vintage materials to developing a distinctive visual style through years of experimentation—are overshadowed by the instantaneous instrumental powers offered by AI. Its _anticipated_ capabilities are stepping stones en route to the inevitable arrival of god-like AGI, rendering human creative labor not just slow but obsolete. Indeed, the tool metaphor hinges on temporal displacement and AI’s instrumental powers—what it can do now (present), how much better a new model performs compared to a previous model (past), and how much more enhanced a model will become (future). The present is positioned as a starting point, and a new, visionary alternative is translated into a temporal difference between the flawed past and the optimized future. The present structures and political arrangements—anything that constrains AI adoption—are framed as obstacles that must be overcome to realize the alternative possibilities that this technology will bring. Such a temporal displacement is a companion metaphor that AI companies want us to live by so that they can conveniently continue their operations under the auspices of the yet-to-come technological advancements. The problems of _today_ are rendered secondary to the promise of _tomorrow_. The ontological might-be’s defer current public concerns around the problematic aspects of AI such as its role in visual disinformation, its immense environmental impact, its embedded epistemological biases, and the indifference of its masters towards creative human expression rendered as data points for ever-better, ever-faster, ever-smarter AIs. If public discourse around these critical questions threatens market dominance, a benign, thrilling conversation about the tool’s dazzling performance certainly does not. The question then becomes: what can we do about these ontological and temporal relations, sold by AI companies as an inevitable co-evolution of humans and machines? Indeed, it can feel disempowering to push back against the industry worth billions and a narrative so totalizing. Yet, as economists Daron Acemoglu and Simon Johnson remind us in _Power and Progress_ (2023), technology’s path is not pre-ordained but is shaped by the choices and values of those in power. Resisting the techno-utopian visions of hubristic, messianic tech leaders is not futile, because the direction of progress is still a choice. To reclaim that choice, we need to ground our critique in the tangible and the immediate. To that end, I want to offer three directions: 1. 1. We should develop increased sensitivity to the contextual arrangements, studying rhetorical constructions of technology and asking whose interests these technologies serve. This means treating AI not as a magical force, but as a product of a specific territory, so-to-speak, with specific geopolitical and economic motives. 2. 2. We must resist the tool metaphor. AI is not a tool. It is technology that carries values of a particular group of people, in a particular part of the world. This acknowledgement forces us to confront the ideology built into the very architecture of AI. 3. 3. We should pay close attention to what AI-generated images show and look for visual evidence of how social realities are framed and distorted. We should anchor ourselves in the _now_ and refuse the version of inevitable futures offered by AI elites. 4. 4. We should remember that sometimes there really are just three things worth saying and do just that despite the fact that AI systems have made three-item lists so common that readers might assume that any such list was AI-generated. No datasets were generated or analyzed during the current study. I dedicate this work to Louise Ravelli who continues to inspire my thinking about meaning and context.