Artificial intelligence is no longer only an emerging technology. It is a way of reorganising institutions, work, knowledge, memory and infrastructure. Today’s radar gathers the signals that point to that same idea.

AI enters its institutional phase

Spain has taken a significant step in regulating artificial intelligence with the draft Organic Law for the good use and governance of AI. The rule adapts the European framework to the Spanish context and focuses on human oversight, transparency, the protection of rights and responsible use in public administration.

The news is of interest not only as a legal novelty. It also shows that AI can no longer be understood as an isolated tool: it is beginning to function as institutional infrastructure, as a system of classification, decision, mediation and government.

From an anthropological view, the question is not only whether an AI complies with the rule. The question is what social relations it produces, what harms it can anticipate, who can complain, who understands the procedure and what happens when a technical decision affects a concrete life.

AIthropology Lab key — Regulating AI is not only about setting limits on a technology. It also involves deciding what kind of technical authority we accept, how responsibility is distributed and what capacity people retain to understand, discuss or challenge automated decisions.

Work, automation and power

In the United Kingdom, a TUC-backed report calls for workers to have more of a say in the deployment of AI systems. The debate is not limited to the replacement of jobs: it also affects autonomy, surveillance, collective bargaining and the sharing of benefits.

Automation does not arrive in an empty space. It enters labour relations already marked by inequalities of power. That is why AI at work should not be analysed only through productivity, but through the real capacity of people to intervene in how their own activity is organised.

AIthropology Lab key — A technology can present itself as neutral and, at the same time, reorganise very concrete hierarchies. It can decide which task is worth something, which person performs, which behaviour is considered normal and how much room is left to resist.

Education, orality and presence

In Cultural Anthropology, Alyssa Paredes defends the value of the oral exam in the age of ChatGPT. The proposal is not to reject AI, but to remember that learning also involves conversation, improvisation, presence and relationship.

The university is leaving behind the simple question of plagiarism. The underlying issue is what forms of learning remain meaningful when writing, summarising or answering can be partly delegated to a machine.

AIthropology Lab key — The oral exam appears here not as nostalgia, but as a reminder that learning is not only producing a correct answer. To learn is also to listen, to doubt, to argue and to sustain an intellectual relationship.

The anthropology of AI as an emerging field

A recent piece by Matt Artz in General Anthropology directly raises the need for an “AI Anthropology”: an anthropology of, with and for artificial intelligence.

The proposal fits an increasingly clear intuition. AI is not only a technical object: it is a cultural practice, a social infrastructure and a field of dispute over knowledge. It is not only that anthropology should criticise AI from the outside, but that it can also study it as a cultural phenomenon, use it carefully in research and bring an anthropological view to its design, regulation and public evaluation.

AIthropology Lab key — The question is no longer only how we use AI, but how we study it, how it transforms us as researchers and what anthropology can contribute to its design, regulation and critique.

Authority, knowledge and algorithmic spectacle

Other recent anthropological debates warn about the academic authority of AI. Sometimes it appears as a guru, as if it offered clarity and guidance. At other times it functions as a conjurer, producing brilliant answers that hide where they come from, what they simplify and what kind of authority they represent.

The problem is not only whether AI gets things right or wrong. The problem is how it produces trust, how it manufactures the appearance of knowledge and how it transforms the relationship between authorship, interpretation and responsibility.

AIthropology Lab key — Generative AI does not only answer questions: it also produces styles of authority. That is why it needs to be analysed as a cultural technology of credibility.

Cultural heritage and conversational archives

A recent piece of work on RAG and cultural collections explores the move from open archives towards local chatbots linked to specific collections. The idea is relevant for museums, archives, the digital humanities and heritage.

Heritage AI should not function as a generalist oracle. It can be thought of as a situated mediation between institutions, curators, documents, public memory and communities. Another study on AI-assisted translation of rock-art texts shows that glossaries improve terminological accuracy. In cultural heritage, mistranslation is not a minor error: it can alter the public understanding of a practice, an object or a tradition.

AIthropology Lab key — An archive is not only an accumulation of data. It is a way of ordering the past, authorising narratives, preserving absences and deciding what may be consulted, translated or reinterpreted.

The cloud has ground, cables and a bill

The energy debate is also gaining weight. Recent reports on data centres in Ireland point to the impact of the growth of these infrastructures on electricity consumption and household bills. At the same time, TSMC acknowledges that AI’s energy demand is forcing a rethink of chip design, prioritising efficiency as well as power.

AI does not float in the cloud. It lives in data centres, electricity grids, chips, territories, water, cables and political decisions about who pays the cost of its expansion.

AIthropology Lab key — The infrastructure of AI is also a social question: it has ground, consumption, energy dependence, territorial impacts and unequally distributed costs.

Closing

Today’s radar points to one and the same idea. Artificial intelligence is no longer only an emerging technology: it is a way of reorganising institutions, work, knowledge, memory and infrastructure. Looking at it anthropologically makes it possible to shift the question. It is not enough to ask what AI can do. We must ask what social world it is helping to build.

Sources