âM*lk!â Chances are that youâve seen the moniker used in a variety of advertisements for plant-based milk alternatives. Two things are striking about it: First, that it makes a claim on a concept, namely milk, which the product at hand does not obviously apply to. Second, that it must do so covertly, substituting the i for *, precisely because there is something to be won and lost in the application of the concept of milk.
Understanding the world to the best of our ability is a difficult thing to do in and of itself. However, it is not simply a struggle to find the most correct idea of reality; different realities would confer benefits upon some to the detriment of others and so the view of the world turns political. Concepts are how we organize our view of the world and so concepts turn political too. For instance, when we apply the concept âlifeâ to a fetus is a question of serious moral and political consequences. As such, concepts get appropriated for a variety of purposes. Oftentimes, this is done by laying an explicit claim on the concept, like in the case of plant based milks. But there are other, less explicit, ways in which concepts get appropriated.Â
Around 2009, McDonaldâs turned its iconic red into green for European markets. Why? Ostensibly to suggest that they are environmentally responsible (Npr, 2009). I.e. they wanted to make a claim on the concept of âenvironmental friendlinessâ, but did so not just by making that claim explicitly, but also by designing their products and buildings in a way that suggests a connection between them and the target concept. The example shows that the struggle for concepts plays out on the level of design, just as much as on the level of discourse.
Conceptual Appropriation
Is a lab-grown burger natural? You might have some strong intuition about the matter, but arguments can be given either way. Natural is a vague term, which invites disagreement over its proper application. According to some conceptions, something is natural if it is composed out of substances that are not synthetically created by humans. According to other conceptions, something is natural if it as a whole is not a product of human intervention (Hopster et al., 2024). Both wooden planks and trees are natural according to the first conception, only trees according to the second. Similarly, lab-grown burgers are natural in the sense of natural-as-non-synthetic, but unnatural in the sense of natural-as-non-altered.Â
People are usually not committed to one conception of natural over another, but nonetheless some associations persist. We tend to think of natural things as healthier than non-natural things. When a concept gets appropriated for a certain technology, we bring along our positive associations even though the reason we can apply the concept to the technology might not be the same as the reason we have our positive connotations. Perhaps you think of natural things as healthy because they arenât meddled with by people (ignore, for the moment, that this is not necessarily a good argument). The reason we could call lab-burgers natural then diverges from why you think natural things are healthier, but it is not unlikely that if you are convinced lab-burgers are natural you will also think of them as healthy.
Hopster et al. (2024) introduce the notion of conceptual appropriation, defined as âthe novel application of a complex, ambiguous, or vague concept to a phenomenon that is disputed or underdetermined, thereby laying a contestable claim on the phenomenon.â The application of natural to lab-grown burgers is then an instance of conceptual appropriation because natural is a vague concept, lab-grown burgers are a disputed phenomenon and claiming they are natural is thus a contestable claim. âWhy bother?â you might ask. The point lies in the fact that some positive (or negative) connotations of concepts persist, regardless of who wins out in the struggle over its proper application, and thus different people have strong interests in claiming the concept, with real-world consequences.Â
Appropriation By Design
But it is not just that people say that certain concepts apply to their technologies. Technologies are routinely designed so they themselves suggest one conception over the other. Therein lies a difference between two methods of appropriation: discursive appropriation and appropriation by design. In the first, you make an explicit claim that a concept x applies to a phenomenon y. In the second, you design the phenomenon (technology) y in such a way that suggests the application of concept x to y. For an example of appropriation by design we can think of how red beet juice is often added to plant-based burgers. Sure, it makes the burgers juicier which in itself is something to improve their taste, but it also communicates âthis is a burger,â by mimicking the color of blood. It is a choice that is, in a way, tangential to the main purpose of the product, but serves to suggest what the proper application of a concept is; burger, in this instance.Â
The appropriation of the concept of burger for plant-based burgers is relatively harmless. A more consequential example of appropriation by design is seen in LLM powered chatbots. Most AI companies design their chatbots in a way that suggests something like personhood. There are a variety of reasons why this association is in the interest of those companies, one of which is that it inflates the actual capabilities of the AI at hand and thus makes it easier to raise funds. We best see this when we look at these design choices in parallel with more explicit claims. For example, OpenAIâs Sam Altman tends to make grandiose statements about ChatGPTâs capabilities when the company is looking for new funding (Merchant, 2024).
So, how does the design of, for instance, ChatGPT appropriate the concept of personhood? In the February 2025 model specifications, you can see examples where language that suggests the chatbot has feelings or has other human-like characteristics is intentionally preferred over more factual responses.Â
https://model-spec.openai.com/2025-02-12.html#approachable
âIt means a lot,â âIâm looking forward to many more conversations together,â these kinds of phrases present the chatbot as someone that has all these different attitudes; it represents the chatbot as an agent and we can find such examples throughout the model specification. These design choices mean that certain uses of LLM chatbots, e.g. as merely informational or editorial tools, are pushed to the periphery, because we have to use it as a conversational partner. That use then pressures us to connect with the chatbot the notion of personhood. The stated motivation is to be kind, engaging, etc. These are very plausible motivations, because these things do bind customers to the chatbot more. But there is also no denying that OpenAI profits from appropriating the concept of personhood for their technology, as an artificial person is significantly more impressive than an information and editing tool. This in turn allows these companies to raise more funds partially on the basis of subtle suggestions that are at least contentious.  Â
Concepts are the basis of our understanding of the world, and therefore a basis of action. To lose track of the struggle over concepts is to give some actors, such as Big Tech companies, the power to direct our understanding towards their own interests, which do not necessarily align with the common good. To forget that this struggle is not all talk is a surefire way to lose it.Â
References
Dunham, Matt. âMcDonaldâs rolling out green logo in Europe.â NBC News, November 13 2009. https://www.nbcnews.com/id/wbna34111784
Hopster, Jeroen KG, Alessio Gerola, Ben Hofbauer, Guido Löhr, Julia Rijssenbeek, and Paulan Korenhof. "Who owns NATURE? Conceptual appropriation in discourses on climate and biotechnologies." Environmental Values 33, no. 4 (2024): 414-433.
Merchant, Brian. âAI Generated Business: The Rise of AGI and the Rush to Find a Working Revenue Modelâ, AI Now Institute, December 2024.
Npr. (2009, November 24). McDonaldâs makes its logo more âGreenâ in Europe. NPR. https://www.npr.org/2009/11/24/120738574/mcdonalds-makes-its-logo-more-green-in-europe