A working guide · nine sections · revised October 2026
Sociology in the age of artificial intelligence
Machines now sort job applicants, rank neighbourhoods, write the news and keep people company at night. Sociology has spent two centuries studying exactly the forces those machines are made of. This is a guide to that inheritance, and to what it lets us see.
Every serious argument about artificial intelligence turns out, sooner or later, to be an argument about society. Whether a hiring model is fair depends on what we think a fair labour market looks like. Whether a chatbot that people confide in is good for them depends on what we believe about loneliness, trust and intimacy. Whether automation will hollow out the middle class is a question about class. The technical vocabulary is new. The questions are not.
Sociology is the discipline that took those questions as its subject. It was born in the early nineteenth century, when another wave of machines was rearranging work, family and belief, and its founders were trying to work out what holds a society together when the old ties stop working. They were wrong about a lot of things. But they built tools for seeing patterns no single person can see from where they stand: who gets ahead and why, how rules become habits, how institutions think, how a crowd decides what is true.
This site is an attempt to lay those tools out properly, in enough depth to be useful, and then to turn them on the present. It is not a textbook summary and it is not a hype piece. Where a famous study has failed to replicate, it says so. Where a theory was used to justify something terrible, it says that too. Where sociologists disagree, which is often, the disagreement is shown rather than smoothed over.
How it's organisedContents
- 01Origins: how sociology came to be
From Ibn Khaldun to computational social science. The founders, the schools, the arguments, and a filterable timeline of nearly two hundred years.
- 02Theory and politics
How ideas about society became political programmes — socialism and social democracy, modernisation and dependency, the welfare state, the Third Way, the New Right — and why the same theory often ends up serving both sides.
- 03The studies
Twenty-seven pieces of research that changed what we know, from Durkheim's suicide tables to audits of medical algorithms. Each with its method, its finding, and what has happened to it since.
- 04Events that moved the world
Revolutions, depressions, uprisings and pandemics, read the way sociologists read them — including People Power in 1986, the Arab Spring, and the Hollywood strikes over AI.
- 05The map: how events reverberate
An interactive map of 34 events and how they cause, echo and provoke one another, from the Industrial Revolution to the Gen Z protests, the Philippine flood-control scandal and AI's squeeze on entry-level jobs. Each event ripples outward from society to institutions to everyday life to you.
- 06The algorithmic society
The long centre of the site. Classical theory applied to machine learning; work and automation; bias and classification; truth and the public sphere; intimacy; governance; and AI as a research tool.
- 07Concepts
A searchable glossary of nearly ninety terms, each with who coined it and why it still matters.
- 08Library
Annotated reading, graded from first book to specialist, plus the datasets, archives and associations worth knowing.
A starting pointThree ideas to carry through the site
First: social facts are real. Émile Durkheim's founding claim was that some things — suicide rates, legal codes, the shared sense of what is shameful — exist outside any single individual and push back against us. You can't wish a labour market away. A trained AI model is a social fact in exactly this sense: it is made from the accumulated traces of millions of people, it outlives any of them, and it constrains what each of them can do.
Second: the way things are is not the way they have to be. C. Wright Mills called this the sociological imagination: the habit of connecting a private trouble (I lost my job) to a public issue (an industry automated). Nothing about the current arrangement of technology, work or power is natural. It was built, by people, for reasons, and it can be rebuilt.
Third: classifications act back on the classified. Ian Hacking called it the looping effect. When we sort people into categories, they respond to being sorted, and the categories change them. Credit scores, risk scores, engagement metrics and recommendation feeds are classification machines operating at a scale no census-taker ever dreamed of. Much of what is distinctive about the present comes down to this.
The question is never just whether a machine is accurate. It is: accurate about what, for whom, decided by whom, and what happens to the people it is accurate about.