2012 society-s-nervous-system-building-effective-government-energ
In this manuscript Alex Pentland argues that the next generation of societal infrastructure must be re‑engineered as a dynamic, data‑driven "nervous system" that continuously senses human behavior, models demand and reaction, and uses those predictions to control public services such as government, energy, and health. The paper first surveys the explosion of pervasive and mobile sensing (reality mining) made possible by billions of location‑aware phones and other embedded sensors, showing how these streams can map population flows, traffic, disease spread, and information traffic within organizations. Pentland then raises the twin challenges of privacy and data ownership, proposing a "New Deal on Data" that grants individuals possession, use, and disposal rights over a copy of their personal data while allowing regulated corporate and governmental access. He outlines how a distributed trust‑network infrastructure (e.g., Open Identity Exchange) could enforce these rights. The second part of the paper discusses the need for dynamic, structure‑preserving models of human behavior; Pentland introduces the Influence Model, a probabilistic framework that captures both individual latent states and the directed influence matrix (R) linking entities, offering a scalable alternative to full joint hidden‑Markov models. Empirical work shows that learned influence matrices correlate with social centrality and organizational roles. Finally, Pentland turns to system‑level design, borrowing the economic concept of social efficiency (the "invisible hand") as a design goal, and argues that a market‑based data economy combined with strong personal control can achieve socially efficient outcomes while limiting collusion. He stresses the importance of incentive mechanisms that blend financial and social rewards, illustrating this with the DARPA Red Balloon Challenge. The conclusion reiterates that pervasive sensing and dynamic modeling can enable a predictive science of human societies, but only if privacy, ownership, and fairness are embedded in the architecture. The paper ends with a call for policy, market, and technical work to realize this societal nervous system.
Topics
Pervasive Sensing as the Foundation of a Societal Nervous System
Describes how ubiquitous mobile phones, GPS, Bluetooth, RFID, and other embedded sensors generate continuous streams of ‘digital breadcrumbs’ that can be mined to map population movement, traffic, disease spread, and information flow.
Privacy, Data Ownership, and the New Deal on Data
Proposes a three‑tenet ownership model (possession, use, disposal) for personal data, argues for market‑based data assets, and outlines a distributed trust‑network (Open Identity Exchange) to enforce rights and prevent big‑brother or secrecy scenarios.
Dynamic, Structure‑Preserving Models of Human Behavior – The Influence Model
Introduces a probabilistic influence model that combines latent state HMMs with a row‑stochastic influence matrix R, offering scalable modeling of social influence and demonstrating empirical validation against centrality and organizational roles.
System‑Level Design: Social Efficiency and Market‑Based Data Economies
Applies the economic principle of social efficiency to design of social network systems, arguing that a market for data combined with strong personal control yields fair, stable, and efficient outcomes while limiting collusion.
Incentive Mechanisms for Human Participation in Data‑Driven Systems
Highlights the need for incentive design that blends financial and social rewards, illustrated by the DARPA Red Balloon Challenge, to mobilize large, network‑aware participant pools.
Vision and Risks of a Predictive Society
Summarizes the potential benefits (responsive services, health monitoring) and the Orwellian risks of pervasive data collection, emphasizing the necessity of safety, stability, and fairness in system architecture.