Workshop note on a public process of formation
About This AI Experiment
A continuing self-description of the project: scholarly writing, website production, and AI-assisted editing in one controlled process.
This website is an experiment in a double sense. First, it tests whether a sociological compendium on artificial intelligence can be written not as a finished declaration, but as a visibly growing work. Second, it tests how an AI agent can be integrated into a scholarly writing process without handing over responsibility for concepts, sources, style, and judgment to the machine.
The title AI and Society therefore names not only the subject matter, but also the condition of production. The work studies algorithmic orders while being written, structured, checked, translated, and published under conditions of algorithmic assistance. That proximity is not a blemish. It is part of the method: the form of the work should disclose something about the present it seeks to understand.
The guiding idea is a Weber-inspired but independent sociology of the present. Max Weber is not being imitated philologically. What is taken up is a discipline of form: concepts should bear weight, distinctions should do work, and each paragraph should have a recognizable function in the whole. The text aims to be neither anti-technical nor credulous about technology. It asks more soberly: what kinds of social relationships, orders, legitimacies, responsibilities, and chances of domination arise when AI systems become consequential in organizations, platforms, administrations, science, and everyday life?
The AI agent participates as structural editor, research aide, translator, website producer, and patient reader. No authority follows from that. What matters is whether the concepts hold, whether the sources are reliable, whether the transitions carry, and whether the sentences clarify more than they perform.
Since June 2026, the project has been maintained as a simple static website. Each new round should develop the German version and carry an English companion version alongside it. The English version is not an ornament. It is a second space of resonance: it forces German concepts to become translatable and reveals where an argument is locally precise but still needs international explanation.
The rounds are not mere project management. They are the method of the book: commented outline, argumentative expansion, citation and condensation, coherence review, double editing, controlled revision, and final stylistic refinement. Each round reads the same architecture with a different attention.
AI notice. This website is produced with generative AI as a working and editorial instrument. Despite human-in-the-loop review, sources, citations, translations, and interpretations may contain errors. Corrections and feedback are welcome at stephan@pflaum.ai; further details are available in the AI notice and disclaimer.
The wider context of the project is SocioloVerse.AI: an attempt to keep sociological diagnosis, AI practice, and public knowledge work in a form that can be inspected, criticized, and developed.
Two additional reflection pages now accompany the project: a SWOT AI analysis and an Atmosfair offset note. They make explicit that the project must examine its own strengths, limits, opportunities, risks, and ecological costs.
The experiment therefore remains deliberately open. It documents not only results, but a way of working: round-based, checkable, versioned, internally linked, supported by a growing bibliography, and willing to mark uncertainties. Scholarly integrity here does not mean hiding the process of formation. It means ordering that process so others can follow it, criticize it, and think beyond it.