Writing on the deliberate design of how organizations and systems sense, reason, decide, and act.https://intelligence-architecture.org/https://intelligence-architecture.org/articles/an-ai-answer-is-not-a-decision/https://intelligence-architecture.org/articles/an-ai-answer-is-not-a-decision/State v. Loomis allowed a COMPAS risk score only a bounded role in sentencing. Luhmann's concepts of decision communication, decision premise, and uncertainty absorption locate the organizational decision around the output.Mon, 20 Jul 2026 00:00:00 GMTfoundationssystemsorganizationsdecision-makinghttps://intelligence-architecture.org/articles/jennings-the-machines-ceiling/https://intelligence-architecture.org/articles/jennings-the-machines-ceiling/Bernard Jennings separates intelligence from computational capacity using three criteria. The proposal is testable and organizationally useful, but neither canonical nor empirically validated, and its ceiling depends on a separate completeness claim.Sat, 18 Jul 2026 00:00:00 GMTfoundationsartificial-intelligencecritiquehttps://intelligence-architecture.org/articles/russell-uncertainty-as-virtue/https://intelligence-architecture.org/articles/russell-uncertainty-as-virtue/The original Off-Switch Game makes human intervention valuable under narrow assumptions. A field study and the Partially Observable Off-Switch Game expose the information, authority, and communication real review still requires.Mon, 13 Jul 2026 00:00:00 GMTfoundationsartificial-intelligencedecision-makinghttps://intelligence-architecture.org/articles/beer-organization-as-controller/https://intelligence-architecture.org/articles/beer-organization-as-controller/A telex network helped Chile coordinate during the October 1972 strike. The Viable System Model and Ashby's law clarify what it regulated — provided the network, the unfinished model, and the Operations Room remain distinct.Tue, 07 Jul 2026 00:00:00 GMTfoundationssystemsorganizationshttps://intelligence-architecture.org/articles/three-architects-of-intelligence/https://intelligence-architecture.org/articles/three-architects-of-intelligence/Beer's Viable System Model, Russell's work on objective uncertainty, and Jennings' recent architectural proposal do not form one theory. As an editorial synthesis, they clarify purpose, deviation, and the right to revise or stop organizational AI.Tue, 30 Jun 2026 00:00:00 GMTfoundationssystemsartificial-intelligencecritiquehttps://intelligence-architecture.org/articles/intelligence-defined-wrong-question/https://intelligence-architecture.org/articles/intelligence-defined-wrong-question/Jennings develops a demanding but as yet empirically untested definition of intelligence. Even a positive verdict under it would confer no right to authorize, override, or decide for an organization.Thu, 18 Jun 2026 00:00:00 GMTcommentaryfoundationscritiquehttps://intelligence-architecture.org/articles/ai-is-not-a-power-drill/https://intelligence-architecture.org/articles/ai-is-not-a-power-drill/The Allegheny Family Screening Tool shows why a model cannot be evaluated in isolation: prediction, policy, roles, and exceptions form the operational system.Tue, 09 Jun 2026 00:00:00 GMTartificial-intelligencefoundationsorganizationshttps://intelligence-architecture.org/articles/trust-is-a-procedure/https://intelligence-architecture.org/articles/trust-is-a-procedure/A plausible AI output cannot certify itself. The Horizon case reveals why provenance, independent review, contestability, and accountable correction must belong to the procedure around it.Thu, 28 May 2026 00:00:00 GMTartificial-intelligenceorganizationsdecision-makinghttps://intelligence-architecture.org/articles/numbered-therefore-true/https://intelligence-architecture.org/articles/numbered-therefore-true/Robodebt shows how an unsupported distribution assumption turned aggregate tax data into an exact individual demand — and why our own playbooks and scores deserve the same test.Tue, 19 May 2026 00:00:00 GMTdesign-theoryfoundationscritiquehttps://intelligence-architecture.org/articles/what-is-intelligence-architecture/https://intelligence-architecture.org/articles/what-is-intelligence-architecture/What Intelligence Architecture means for the connection of human and machine work, why adjacent vocabularies do not quite cover the same question, and why this difficult term can still open useful ground.Fri, 08 May 2026 00:00:00 GMTfoundationsorganizationsdesign-theoryhttps://intelligence-architecture.org/articles/machines-that-reason-with-us/https://intelligence-architecture.org/articles/machines-that-reason-with-us/Wong and colleagues found weak discrimination in a retrospective single-center study of the then-current Epic Sepsis Model. The result is not causal evidence of harm, but it reveals how much organizational power can be attached to a score.Thu, 30 Apr 2026 00:00:00 GMTartificial-intelligencedesign-theorysystemshttps://intelligence-architecture.org/articles/decision-flows-in-institutions/https://intelligence-architecture.org/articles/decision-flows-in-institutions/The three stalls in this essay are an authorial heuristic, not a settled theory. Columbia forces revisions to all three because registration, translation, evidence standards, and authority overlapped.Tue, 21 Apr 2026 00:00:00 GMTorganizationsdecision-makingsystems