AI Orchestration of Human-to-Human Educational Networks

The dominant Paradigm for AI in education treats machine intelligence as a substitute for human instruction — an always-available tutor, interlocutor, or socratic partner that scales without the friction of coordinating real people. This framing accepts scarcity of human attention as a fixed constraint and deploys technology to work around it. But there is a coherent alternative architecture in which that scarcity is itself the problem to be solved, not the premise to be accommodated.

The alternative vision positions AI as an orchestration layer for peer-to-peer and intergenerational knowledge exchange. Any sufficiently large human population contains a staggering combinatorial density of latent educational relationships: people who possess skills, domain knowledge, or craft expertise that others in proximity want to acquire. The coordination problem — matching availability, interest, location, and pedagogical fit across thousands of simultaneous potential dyads — is precisely the class of problem at which machine learning systems excel. This is, in essence, a technically sophisticated realization of Ivan Illich's 'learning webs' from deschooling Society: a network infrastructure that makes the full educational resources of a community legible and navigable without centralizing or Institutionalizing them.

The reason this architecture was not built is not technical but economic. Platforms that mediate human-to-human connection at low margin are structurally less attractive than platforms that substitute for human connection at high margin. The incentive gradient of the technology industry points toward products that are maximally sticky and substitutive — exploiting the human appetite for social and intellectual engagement while externalizing the costs of genuine community infrastructure. The path of Scaffolding was available; the path of replacement was more profitable.