A Critical Vocabulary FOR Artificial Intelligence

The AI industry supplies most of the words we use to describe its technology. It has been delivered to us through analogies of thinking machines, intelligence and parallels to human cognition. Generating new vocabulary is a useful tool for creating clarity beyond the industry’s frame. This page collects and clarifies terminology I’ve used in my writing and thinking about the politics & ideology of AI, particularly large language and diffusion models. If you find these framings helpful, feel free to borrow them (recommended citations are listed after each section; you can find a stable reference to this glossary here).

Myths ~ Systems From Nowhere ~ Radical Intentionalism ~ Hypothetical Images ~ Noise ~ The Age of Noise ~ Stochastic Flocks

“The words and ideas that distort the industry’s understanding of the products it is itself building.”

The Myths of Generative AI

The AI industry is animated by myths. My use of this term, myths, is not a test of whether something is true or false. For Barthes, a myth is an extra layer of meaning added to the signifier. We can say that "baseball is a myth," for example, if we associate it with what it means to be American, or we can say "air travel is a myth" if we associate it with progress and modernity. Often, this is not explicit — we may never say out loud that air travel represents modernity, but we feel it if, for example, we visit a country with badly run airports.

For the AI industry, this means that a handful of ideas serve as a lens for its own understanding of the products it is building. The AI industry has many myths that animate it, such as intelligence, learning, scale, productivity, and emergence.

Myths become concerning when the "extra" associated with them become mistaken for reality. For example, there is nothing much wrong with suggesting a machine "learns" from data. But if we then argue that the machine has the same rights or privileges as a student, we inflate the myth of learning machines to justify things that would make little sense without the myth: for example, arguing that tech companies should be allowed to enroll a language model in a university, record the professor's lectures, and use them as a source of training data to build a competing educational product.

Likewise, the productivity myth suggests that AI will boost productivity. Productivity is certainly a myth: it is assumed to represent economic progress, and a better quality of life. But it is unclear that any productivity associated with large language models will positively influence salaries, employment, or social benefits. One does not need to believe this is false. But as a myth, it is important to disentangle the word "productivity" from these associated meanings, and to ask clarifying critical questions.

Related essays

“The habit of describing AI systems as isolated technical objects, without referencing the decisions of the engineers who build them, which relocates accountability from people to models.”

The System From Nowhere

“The system from nowhere" is a way of talking about AI systems as if they were a spontaneously emerging force rather than a human-engineered product.

It happens when industry leaders, the media, or policymakers refer to the system's actions rather than decisions made by the people who designed the system. Examples of this include describing "rogue AI," as if the machine has taken action of its own (non-existent) volition.

This is not to say that systems are predictable. Rather, it emphasizes their unpredictability as a scapegoat. When OpenAI creates a test of a model optimized to find exploits in software, they are designing a set of conditions and environment in which unexpected things may occur. If they fail to disconnect internet access from that network, they are responsible for the security gap. When the model goes online, as it did in the 2026 Hugging Face hack, the company — and media, and policymakers — directed their concerns to the "rogue model," rather than examining the decisions OpenAI made about how to contain its model within the test environment.

The System From Nowhere occurs elsewhere in policy. Over the years, major policy documents have focused attention on the actions of models as a priority for policy. The UN, for example, has removed any acknowledgement of the people behind artificial intelligence, warning instead that artificial intelligence is developing faster than we can control it. The error arises from drawing a boundary around these systems at the technical level: defining the system entirely by what happens within the machines, rather than the people who build and deploy the machines.

This framing erases accountability — bad decisions can be blamed on "rogue models" — but it also erases materiality: the "system from nowhere" often severs the model from the data centers that are required for them to run, the bureaucratic arrangements surrounding their deployment. It also strips away the human labor that is less powerful, but still present, in building these systems: people like the data annotators paid to improve the training data; or people like you and me whose images and text have been incorporated as the raw material for the model's training.

The concept borrows on critiques of Nagels “view from nowhere” via Jay Rosen and Donna Haraway, with a lineage discussed in the primary essay.

"The system from nowhere invites us to imagine AI without people. And because the people are doing so much, we have to fill in that absence, placing the model in the position of the God-trick, or an oracle, or an inevitable, emergent superintelligence. It's a fantasy, an invitation to streamline the world's contradictions into a coherent story."

Related essays

Radical Intentionalism

Radical intentionalism insists upon interpretations of AI that center “beliefs,” “thoughts,” and “intentions” from the model as the exclusive lens of analysis, while also rejecting the value of questions about design and deployment decisions.

The system from nowhere is a consequence of what I call “radical intentionalism” — after what Daniel Dennett calls the intentional stance, which asks whether treating an object, such as an LLM, as something with beliefs is useful for predicting what it does. Dennett also describes a design stance, in which predictions are drawn from knowledge of the purpose of the system: asking not “what did the model think it was doing,” but “what was the model designed to be doing?”

It seems inevitable to use the intentional stance with complex systems: “what is the model ‘trying’ to do” can be a genuinely useful question. But its designers must not stop there. The design stance is equally valuable: it tells you where the “trying” actually came from, and asks what the designers were trying to get the model to do. The two operate in tandem. The radical intentionalist stance, currently in vogue across the tech industry, insists that the intentional stance is the only explanation worth pursuing.

From my 9/2026 piece in Tech Policy Press: “In industry rhetoric, a radical intentionalist stance dominates any interpretation of model behavior, while the explanatory power of the design stance is rejected outright. This is a strange and dangerous relationship for designers to have with a product they are literally designing. Predicting the actions of one’s product from the intentional stance alone carries political consequences for the rest of us. It proposes a system from nowhere: a technical system that exists only as a technical object, without acknowledging the decisions that built it. Yet for an industry and a set of AI risk critics who argue that its products are “grown, not built,” it shoos us away from the rot beneath the flower box.”

Crucially, I believe both the soft-intentionalist stance and the design stance have a role to play, and I do not reject the benefits of intentional readings. The key is the modifier, “radical,” which speaks to the insistence that only the intentional reading is acceptable. The ideal is a flexible stance, where intent can be used to describe and anticipate the behavior while the design stance illuminates the precise mechanisms that might shift this behavior.

Related Essays

“A hypothetical image is an image of something that exists only as a statistical likelihood: the output of a latent diffusion model is better understood as a data visualization constrained by its training data than as a picture of the world.”

Hypothetical Images:
AI Image as Infographic

A hypothetical image is an image of something that exists only as a statistical probability. While images made by latent diffusion models are images, they are better understood as data visualizations, or infographics. What they depict is constrained by the images used in the training data. We can think of the image, then, in the mathematical sense of a hypothesis: just as data is used to understand patterns in a line graph or pie chart, the image itself is a result of a prediction based on the data: a prediction of what an image associated with your prompt would look like, based on the data available. Much like a chart plotting numbers on a line, its “predictive” power is anchored entirely in the past.

From How to Read an AI Image, which tests a methodology for “reading” AI images by examining a series of “hypothetical images” of people kissing. We find, for example, that “kissing” prompts always produced heterosexual pairing and pairs with similar skin tones. We also ask: why do these images look so odd?

This data sets the boundaries of what can be represented, and we see this in the data and when we explore the mechanisms within the models. It's important to clarify that this data is not re-created in the resulting image. Instead, all kinds of mechanisms evaluate the image as it moves from its starting point — noise and blur — into something with the resolution of a photograph. The model is designed to produce a series of automated decisions about what an image is supposed to look like at all.

The "hypothetical" image is therefore a way in to ask the questions we would ask of any data visualization: what is an image supposed to look like? We can look at an AI generated image and ask: what data and methods informed this hypothesis? Designers of the system had to ask this too. They embedded their idea of what a “successful” image looks like into the system, so the model could make those decisions, at scale, in their absence. What were those decisions? What assumptions are baked into those mechanics?

Related essays

  • Primary: How to Read an AI Image. Web version. Preferred citation and peer-reviewed version: Salvaggio, E. (2023). How to read an AI image: Toward a media studies methodology for the analysis of synthetic images. https://doi.org/10.25969/MEDIAREP/22328

  • The Hypothetical Image. Argues that the culture and ideology of data surveillance and statistics permeates the images made by generative AI in ways that reduce and neutralize the subjects it trains on. Salvaggio, E. (2023, October 29). The hypothetical image: The aestheticization of algorithmic ideologies (Version 1). Cybernetic Forests. https://doi.org/10.5281/zenodo.22036195

  • The Market in the Model (pre-print). In this technical and ideological audit of the original latent diffusion model, I examine the mechanisms that produce an image and the logic embedded into each step. (A condensed introduction to the paper is here.) Salvaggio, E. (2026). The market in the model: Latent diffusion as neural economy. arXiv [cs.CY]. https://doi.org/10.48550/arXiv.2606.19151

  • From AI Photograph to Hypothetical Image. Argues for the term "hypothetical image" and the critical lens it affords. Salvaggio, E. (2026). Hypothetical images: AI photographs. In The need to rename tech (pp. 171–187). Springer Nature Switzerland.

  • On the practice of AI image making and its impact on personal memory:

“In generative image models, noise is the literal starting material: a field of randomly colored pixels that the system converts into a legible picture by referencing, step by step, what an image is supposed to look like.”

Noise

Noise is literally the technical foundation of generative AI systems, particularly diffusion models used for producing images, video, and sound. Blur and scattered, randomly colored pixels are restructured by the model to arrive at something resembling a concept associated with your prompt. In my work on images, I shift the emphasis from the final output of the system to the seed. This allows me to trace the system's decisions in eradicating noise and replacing it with an artificially generated signal.

From Pollen Series, an image from a bespoke image model trained on public domain images of pollen and failed noise prompts.

I consider the tension that the model is therefore incapable of producing images of noise, because they are designed to strip noise out of the image. The earliest latent diffusion models could not create "noise" prompts, falling into strange swirls of color that had no correspondence to the training data. My perspective is that noise represents contingency, chance and the unpredictable wilderness of the world, which must be stripped away from the image to produce a legible image. Yet, as these images circulate, they become noise themselves (as in the "AI slop" that dominates our social media feed, but also the algorithmically amplified signals that mediate our information).

My artistic practice incorporates artifacts produced by diffusion models when asked to produce images of noise and failing. These images, which do not reference the training data, stand in for the uncomputable: though of course computer systems can produce images of noise, the failure of the diffusion model to do so offers a glimpse into the inherent structure of the model. The contradiction of imposing structure into a prompt designed to remove it points to the human desire to remove the noise of the world, but noise insists on carrying entropy into our systems nonetheless.

Related essays

  • The Market in the Model. Examines the original latent diffusion model to examine the mechanisms that produce an image, asking about the logic and function of each step.

  • A friendly, condensed introduction to the paper is here.

  • Full paper:

“The natural outcome of the information age, where the problem is no longer finding information but filtering it out — and where power isn’t centered in the strength of a signal, but control of the filter.”

The Age of Noise

The age of noise suggests the information age is over. Where information was once scarce, today it's overwhelming. Awash in information, we struggle to understand our world or make decisions. Generative AI (large language and diffusion models) has arrived to fill this gap between information overload, decision paralysis, and the pressure to optimize: it will sort it all out on our behalf — for a monthly fee.

Diffusion models (the tech behind generated images) are built on noise on a genuine technical level: it's a machine that takes the images we made for each other and dissolves them into fields of arbitrary pixels. It then walks the destruction backward, based on a single internally generated image of random static, and adjusts them toward the constraints of our prompts. This is also a social sequence: dissolve communication into a sea of noise, and then algorithmically reconstruct something that resembles communication.

The distance between those points fascinates me. What gets dissolved is context. An archive is curated, contested, and answerable for its gaps; a dataset is an archive with human relationships stripped out and reassembled by different means. What shape does that take?

I work inside that tension rather than at a distance. Through a critical technical practice, I work with AI systems to explore the gaps they instantiate, the accountability they displace, and the intersection of our cultural world with the machinic interpretation of how meaning means. Through films, installations, exhibitions and creative inquiry, I work with and against the technical and social noise of AI to make sense of it, often by misusing the systems until they fail in revealing ways.

Related talks

Stochastic Flocks

“Stochastic parrots in a stack: systems whose outputs become inputs to other systems, generating and modifying text, code, and images that get passed on and expanded again. The problem is the pandemonium: separately optimized systems competing or reinforcing each other to produce cascading results.”

With the rise of "agentic reasoning models," systems are more complex and even more inscrutable, and the temptation to attribute intelligence — that is, thoughtful intent or volition — to the text they produce is even stronger. This has lead many to dismiss the “stochastic parrots” analogy as reductive. I propose that the analogy remains essentially correct, but that as systems have scaled, so should the analogy.

Stochastic parrot has become wrongly associated with a reductionist perspective on large language models and their capabilities. In their foundational 2021 paper, Emily M. Bender, Timnit Gebru, Angelina McMillan-Major and Margaret Mitchell described LLMs as stochastic parrots — systems that "parrot" statistically likely patterns derived from training data. Stochastic implied that these models had some room for variation, but were always constrained by being anchored to the training data. Parrot referred to the fact that the language was passed through, rather than ever being internalized or “understood.”

The paper was written in 2021. Much has changed about language models since, but this all remains true. In 2025, a wave of models, dubbed "agentic" "reasoning" models, were introduced to the public. These newer models stacked various types of systems on top of each other to produce longer sequences of text, and sequences of text that took on the specific flavor of language that people use when solving problems. But as I have written elsewhere: "The language does not emerge from that reasoning, it is the reasoning."

Agentic systems stack these parrots into interacting outputs — what I call a stochastic flock. (Appropriately, the plural noun for a flock of parrots is a pandemonium.) Stochastic flocks describe what industry researchers sometimes refer to as "artificial hivemind," but strips away the illusion of mind that suggests volition or intent. With multiple instances of agents operating between models, or stacks of internal processes behind a single interface, multiple stochastic processes have been tied together to create a more complex system.

In 2019, while working with LLMs in my professional capacity, I wrote about what OpenAI's GPT-2 model suggested about how we would relate to the language produced by machines, and "the challenging work of understanding how human minds respond to generated text; how we begin to read, think, and process what it's telling us, and how to develop 'filters' that protect us from ascribing a humanlike intent." Nothing in the architecture of language models has shifted this view, even as they become more capable of producing complex speech, code, and automate tasks.

The stochastic flock does not deny that models have become more complex. Instead, I use it to distinguish the capacity to produce complex language from the complex understanding of language. We might even argue: "language is a myth," a term with many associations in urgent need of untangling from thought. Yes, models can do things with language, but that language is not evidence of thought. It does not mean they are stupid, merely that they do not think. They are erudite zombies: “they produce text influenced by training data, optimization during pre- and post-training, and the content of the prompt.” In the design and deployment of these systems, language is treated as if it were thought. Failures occur when tasks that require thought are filled in by language as if they did the same work.

A common misread of my perspective is that it says there is nothing to worry about. That is not the implication. Rather, I propose it to reframe the nature of our concerns: not to worry about a superintelligence that becomes smarter than us, but about the risks involved in a language and coding machine that does not think. The belief in the model’s capacity for thought smuggles capacities that do not belong to it; we begin to address abstractions such as the model’s beliefs or intentions. It does not deny the problems or risks of unconstrained model development; it denies the interpretation of their causes.

Agentic systems are more difficult to understand and parse, and so it is more tempting to ascribe a kind of volition to the things that emerge from them. At the heart of critical AI literacy is to understand that language production does not imply intelligence, no matter how fluent the systems become. So some of my theory work is about that distinction: what does it mean to write without thinking, and to read words produced by an absence of thought?

Related essays and talks

 

Stable citable version of this glossary:
Salvaggio, E. “A Critical Vocabulary for Artificial Intelligence: Terms for the Politics and Ideology of Large Language and Diffusion Models”. Version 1.0, Cybernetic Forests, 10 Sept. 2026, https://doi.org/10.5281/zenodo.22692687.