Distribution scales.
Integration does not.

Digital systems have progressively solved the problem of information distribution — while leaving the problem of knowledge integration structurally underdesigned.

Across half a century, the dominant trajectory of digital infrastructure has been the progressive reduction of distribution cost: networks have become denser, faster, more personalised, and more global. The implicit assumption that integration would emerge as a by-product of distribution has not held.

What has emerged instead is a digital society in which the supply of information vastly outpaces the institutional, cognitive, and computational capacities required to integrate that information into stable, contextual, and collectively held knowledge.

Integration
underdesign.

The mismatch between distribution and integration is not a contingent failure of any particular platform. It is a design property of the systems through which digital society conducts itself. Recommender systems, social platforms, generative models, and decentralised protocols are optimised — explicitly or implicitly — for throughput, engagement, and amplification.

None of these objectives reward the slow, costly, contextual work of integration: reconciling claims, stabilising meanings, situating knowledge in shared histories, or maintaining the conditions under which distributed information can be turned into something a community can act on.

I call this absence integration underdesign. It is the central object of the programme.

Network-induced
fragmentation.

The mechanism through which integration underdesign manifests is network-induced fragmentation: a systematic erosion of the structural conditions of knowledge integration as networks densify, accelerate, and personalise.

Fragmentation is not a downstream consequence of bad content or bad actors. It is a property of network architecture under conditions of high distribution and low integrative incentive. Three vectors are central:

Density — as the number of available pathways grows, the coherence of any individual pathway declines. Speed — as flows accelerate, the temporal stability of knowledge structures degrades. Personalisation — as exposure narrows, the overlap of context required for shared knowledge thins.

Three research
fields.

The programme is organised across three substantive fields. Each examines a distinct site at which network-induced fragmentation is observable, measurable, and consequential — and at which the broader concerns of management and innovation dynamics are at stake.

Current conference
contributions.

Two accepted contributions extending the Communication & AI Systems field. Both develop the integrative architecture argument toward the design of human–AI interaction.

What the programme
produces.

The aim of the programme is the development of a theoretical and empirical account of integrative architecture — the principles under which networks, platforms, and AI systems could be designed and governed so that distribution and integration scale together.

Outputs include peer-reviewed publications across the three fields, working papers and essays articulating the framework, applied collaborations with institutions facing concrete integration problems, and teaching at the intersection of network science, management, and innovation.

See publications by trajectory →