Governing AI Outputs in a Fully Sensed Society: Is Output Governance Sufficient for Social Acceptance?

Advances in sensing, the Internet of Things (IoT), and artificial intelligence (AI) are moving smart cities toward a Fully Sensed Society (FSS), in which human behavior and states can be continuously inferred. This paper presents the Fully Monitored Public Space (FMPS) as a scaled-down model of the FSS and a smart-city testbed, and examines two questions. First, to what extent can unauthorized external AI outputs be structurally deterred? The previously proposed Verifiable Record of AI Output (VRAIO) binds outputs to Rules, an independent Recorder, tamper-resistant records, audits, and sanctions, making deviations verifiable and attributable. The Recorder, however, checks only the consistency of declared metadata with the Rules, not the output content; false declarations are deterred because unannounced audits and sanctions make their expected benefit smaller than the expected loss upon detection. Second, if such trustworthiness is secured, does it lead to social acceptance of being continuously known by AI? We propose a staged and reversible evaluation framework for the FMPS that distinguishes acceptance grounded in the experience of “no actual harm” and of “benefits” (personal and social) from resignation accompanied by chilling effects. This paper does not answer the question; it presents a method for asking it empirically.