Manuel Castells on AI & the Programmable Network Society

Teaser

He’d frame AI inside the network society: power flows through programmable networks and codes of communication; identities resist and reconfigure, yet infrastructure quietly steers whose signals travel furthest. To keep agency alive, we must contest who programs the programs, and how identities gain routing privileges in the space of flows (Castells 2010; 2009).

Methods window

Assessment target: BA Sociology (7th semester) — Goal grade: 1.3 (Sehr gut).
Approach. Conceptual reconstruction of Castells’ core motifs—space of flows vs. space of places, programmability, communication power, and identity (legitimizing / resistance / project)—then application to AI search, recommenders, and model-assisted coordination.
Theory anchors. Castells (2010; 2009; 2015; 2001).
Scope. Public-facing AI in media, education, and civic tech; illustrative examples (no proprietary data).
Quality & transparency. APA short style in text; full list (publisher-first links) below; disclosure + check log at the end. Key sources: Castells’ trilogy 2nd ed., Communication Power, and Networks of Outrage and Hope. (wiley.com)

Close-Reading Box: Three Castellsian anchors (no page numbers)

Programmable networks

Networks are programmable: their goals and codes can be reconfigured by those who control infrastructure and protocols (Castells 2010; 2009). AI extends programmability from transport of signals to shaping of meanings (ranking, labeling, generation).

Communication power

Power is exercised by coding/decoding communication flows—framing, connecting, and amplifying messages across media networks (Castells 2009). AI intensifies this via recommendation, synthetic voice, and content policy “grammars.”

Identities in the network society

Actors assemble resistance and project identities that travel through networks (Castells 2010; 2015). But reach is infrastructural: identities gain power when they secure bandwidth, interoperability, and visibility.

Evidence block — Reading AI through Castells

• From access to routing. The AI question is less “who can speak?” than whose packets get priority—which prompts, datasets, and sources are routable at agenda-setting moments (Castells 2010).
• Codes as institutions. Safety policies, ranking rules, and provenance labels act like communication law in code; they distribute attention rights (Castells 2009).
• Space of flows vs. places. Local publics can win voice if they bridge places into flows (schools, libraries, city hubs acting as “signal boosters”)—otherwise they remain edge communities (Castells 2001).

Mini-Meta (2010–2025): What Castells adds now

Across research on platforms and movements, three lessons recur: (1) Infrastructure is political—API gates, app stores, and data pipelines steer visibility; (2) Programmability migrates from transport to meaning-making (search, generative fill, moderation); (3) Identity work succeeds when it couples narrative to infrastructure (toolchains, distribution, translation). Castells’ twist: treat AI as a layer inside communication power, not an external disruptor (Castells 2009; 2015).

Practice heuristics (testable rules)

1. Expose the code path: For each AI output, show source → transform → policy gates (human-readable).
2. Programmability budgets: Reserve % of ranking for non-trained-on-you sources and minority languages/regions.
3. Identity-to-infrastructure bridges: Fund translation + mirroring so resistance/project identities become routable beyond their home place.
4. Interoperable appeals: Standardize notice-reason-appeal across tools so signals can move even when blocked locally.
5. Edge resilience: Cache critical civic content at the edges (schools/libraries) to withstand central outages or de-amplification.

Counterpoints (and why Castells still matters)

• Platform studies (van Dijck): Platforms rewire connectivity through corporate ecosystem logics; not all networks are open networks. Castells helps parse how codes become power inside those ecosystems (van Dijck 2013).
• Data colonialism (Couldry & Mejias): Beyond networks, extraction structures the field. Castells’ communication power complements this by showing how extraction gets narrated and normalized in flows (Couldry & Mejias 2019).

From hypotheses to measures

• H1 (Programmability & reach). Tighter central control of APIs/models increases cross-user similarity of top-N exposure and reduces peripheral source reach.
  Measures: overlap of user top-N; k-core distance of surfaced sources; API outage/price shocks → visibility shifts.
• H2 (Identity + infrastructure). When movements add translation + mirroring, their cross-language cross-region diffusion grows despite ranking changes.
  Measures: share of first-time regions reached; multilingual engagement ratio.
• H3 (Appeal portability). Interoperable appeals reduce wrongful downranking/removal persistence.
  Measures: time-to-reinstatement; cross-platform carry-over rate of successful appeals.

Sociology Brain Teasers

• Where do your messages die: at access, routing, or decoding?
• Which code rule (ranking/safety) most shapes your field—and how would you rewrite it in one sentence?
• If you had a 10% programmability budget, what would you allocate it to?
• Name one edge institution (school, library, union) you could equip as a signal booster.
• Draft a portable appeal you’d accept across platforms.

Transparency & AI disclosure

Co-produced with an AI assistant (GPT-5 Thinking). Human lead: Dr. Stephan Pflaum (LMU Career Service). Workflow: outline → close reading → drafting → counterpoints → hypotheses/measures → APA/QA. No personal data used. Limits: interpretive essay; models can err; conjectures are flagged as such. Contact: contact@sociology-of-ai.com. Post_id: sai-2025-11-07-castells.

Check log

• Teaser ✓ • Methods ✓ • Close-reading (3) ✓ • Counterpoints ✓
• Heuristics (5) ✓ • Hypotheses→Measures ✓ • Brain Teasers (5) ✓
• APA short style in text (no page numbers) ✓
• Assessment target present ✓ • Disclosure ✓ • Header image 4:3 + alt text ✓

---

Literature (APA, publisher-first links)

• Castells, M. (2010). The rise of the network society (2nd ed.). Oxford, UK: Wiley-Blackwell. (wiley.com)
• Castells, M. (2010). The power of identity (2nd ed.). Oxford, UK: Wiley-Blackwell. (wiley.com)
• Castells, M. (2010). End of millennium (2nd ed.). Oxford, UK: Wiley-Blackwell. (wiley.com)
• Castells, M. (2009). Communication power. Oxford, UK: Oxford University Press. (Oxford University Press)
• Castells, M. (2015). Networks of outrage and hope: Social movements in the Internet age (2nd ed.). Cambridge, UK: Polity. (politybooks.com)
• Castells, M. (2001). The Internet Galaxy: Reflections on the Internet, Business, and Society. Oxford, UK: Oxford University Press. (OUP Academic)

Further reading

• van Dijck, J. (2013). The culture of connectivity: A critical history of social media. Oxford, UK: Oxford University Press. (OUP Academic)
• Couldry, N., & Mejias, U. A. (2019). The costs of connection: How data are colonizing human life and appropriating it for capitalism. Stanford, CA: Stanford University Press. (Stanford University Press)

Manuel Castells as a Forerunner of AI-Era Sociology: What He Got Right, Partly Right, and What Doesn’t Fit

Header image (for Gutenberg cover block)

Alt text: “Abstract 4:3 composition—interlaced signal lines forming hubs and bridges; a few paths glow to suggest programmable routing in the network society.”