<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Intelligence Architecture</title><description>Writing on the deliberate design of how organizations and systems sense, reason, decide, and act.</description><link>https://intelligence-architecture.org/</link><item><title>The Business Value of Critical Thinking When AI Does the Work</title><link>https://intelligence-architecture.org/articles/business-value-of-critical-thinking/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/business-value-of-critical-thinking/</guid><description>What is critical thinking worth when AI agents do the work? Why Intelligence Architecture must develop employees’ capabilities, and how organizations can assess the business value.</description><pubDate>Fri, 06 Nov 2026 00:00:00 GMT</pubDate><category>intelligence-architecture</category><category>organizations</category><category>artificial-intelligence</category></item><item><title>AI and the Unwritten Rules of an Organization</title><link>https://intelligence-architecture.org/articles/invisible-systems-in-organizations/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/invisible-systems-in-organizations/</guid><description>AI systems know an organization through the information available to them. How can its unwritten rules be recorded without obstructing work or creating new surveillance?</description><pubDate>Thu, 03 Sep 2026 00:00:00 GMT</pubDate><category>intelligence-architecture</category><category>organizations</category><category>decision-making</category><category>foundations</category></item><item><title>How an AI Output Becomes Grounds for a Decision</title><link>https://intelligence-architecture.org/articles/an-ai-answer-is-not-a-decision/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/an-ai-answer-is-not-a-decision/</guid><description>In State v. Loomis, a US court considered a software-generated risk assessment in sentencing. The case shows why rules and assigned responsibilities must constrain the influence of such outputs.</description><pubDate>Mon, 20 Jul 2026 00:00:00 GMT</pubDate><category>foundations</category><category>systems</category><category>organizations</category><category>decision-making</category></item><item><title>Why AlphaZero Does Not Count as Intelligent for Bernard Jennings</title><link>https://intelligence-architecture.org/articles/jennings-the-machines-ceiling/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/jennings-the-machines-ceiling/</guid><description>Why does Bernard Jennings classify AlphaZero, the chess, shogi, and Go system, as non-intelligent? An examination of his definition and proposed ceiling on intelligence.</description><pubDate>Sat, 18 Jul 2026 00:00:00 GMT</pubDate><category>foundations</category><category>artificial-intelligence</category><category>critique</category></item><item><title>Stuart Russell and the Question of Keeping AI under Human Control</title><link>https://intelligence-architecture.org/articles/russell-uncertainty-as-virtue/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/russell-uncertainty-as-virtue/</guid><description>When does an AI system benefit from letting a person decide whether it should act? This article examines Russell&apos;s model and the conditions for human control in the workplace.</description><pubDate>Mon, 13 Jul 2026 00:00:00 GMT</pubDate><category>foundations</category><category>artificial-intelligence</category><category>decision-making</category></item><item><title>Stafford Beer and the Limits of Project Cybersyn</title><link>https://intelligence-architecture.org/articles/beer-organization-as-controller/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/beer-organization-as-controller/</guid><description>Stafford Beer’s Project Cybersyn aimed to improve factory coordination in Chile. What did it achieve, and what does his model of organization contribute to the design of AI deployments?</description><pubDate>Tue, 07 Jul 2026 00:00:00 GMT</pubDate><category>foundations</category><category>systems</category><category>organizations</category></item><item><title>Beer, Russell, and Jennings on What Comes Before Choosing an AI Model</title><link>https://intelligence-architecture.org/articles/three-architects-of-intelligence/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/three-architects-of-intelligence/</guid><description>What must be decided before choosing an AI model? This essay connects Beer’s organizational model, Russell’s work on human control, and Jennings’s provisional definition of intelligence.</description><pubDate>Tue, 30 Jun 2026 00:00:00 GMT</pubDate><category>foundations</category><category>systems</category><category>artificial-intelligence</category><category>critique</category></item><item><title>Bernard Jennings’s Definition of Intelligence and Its Limits for AI Deployment</title><link>https://intelligence-architecture.org/articles/intelligence-defined-wrong-question/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/intelligence-defined-wrong-question/</guid><description>Bernard Jennings defines intelligence through a system’s internal operation. This article examines the untested proposal and explains why intelligence would confer no authority to decide.</description><pubDate>Thu, 18 Jun 2026 00:00:00 GMT</pubDate><category>commentary</category><category>foundations</category><category>critique</category></item><item><title>How a Risk Score Changed Child-Welfare Work</title><link>https://intelligence-architecture.org/articles/ai-is-not-a-power-drill/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/ai-is-not-a-power-drill/</guid><description>Software supports child-welfare decisions in Allegheny County. How did a threshold of 18 become a rule for investigations, and how did it change the work of staff and supervisors?</description><pubDate>Tue, 09 Jun 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><category>foundations</category><category>organizations</category></item><item><title>What Makes an AI Output Worth Trusting</title><link>https://intelligence-architecture.org/articles/trust-is-a-procedure/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/trust-is-a-procedure/</guid><description>The Horizon scandal in British post offices shows the consequences of software outputs that are hard to verify. What follows for evidence, challenge, and responsibility when using AI?</description><pubDate>Thu, 28 May 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><category>organizations</category><category>decision-making</category></item><item><title>Robodebt: How Estimated Income Became a Debt</title><link>https://intelligence-architecture.org/articles/numbered-therefore-true/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/numbered-therefore-true/</guid><description>Australia’s Robodebt system turned an assumption about income over time into grounds for debt claims. What the case reveals about the limits of machine-generated assessments.</description><pubDate>Tue, 19 May 2026 00:00:00 GMT</pubDate><category>design-theory</category><category>foundations</category><category>critique</category></item><item><title>What Is Intelligence Architecture?</title><link>https://intelligence-architecture.org/articles/what-is-intelligence-architecture/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/what-is-intelligence-architecture/</guid><description>Intelligence Architecture examines how people and AI work together in organizations. An introduction to the term, its limits, and the design of tasks and responsibilities.</description><pubDate>Fri, 08 May 2026 00:00:00 GMT</pubDate><category>foundations</category><category>organizations</category><category>design-theory</category></item><item><title>When Hospital Software Is Meant to Warn of Sepsis</title><link>https://intelligence-architecture.org/articles/machines-that-reason-with-us/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/machines-that-reason-with-us/</guid><description>A study of the Epic Sepsis Model estimated missed sepsis cases and the reviews its alerts would have prompted. What do those findings mean for the work of hospital clinicians?</description><pubDate>Thu, 30 Apr 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><category>design-theory</category><category>systems</category></item><item><title>Columbia: How Warnings Lost Their Significance on the Way to a Decision</title><link>https://intelligence-architecture.org/articles/decision-flows-in-institutions/</link><guid isPermaLink="true">https://intelligence-architecture.org/articles/decision-flows-in-institutions/</guid><description>The Columbia investigation shows how evidence of damage lost its significance for mission decisions. What can organizations learn from it when designing AI-supported information flows?</description><pubDate>Tue, 21 Apr 2026 00:00:00 GMT</pubDate><category>organizations</category><category>decision-making</category><category>systems</category></item></channel></rss>