From a61dabc8af8fb95c9bb5622ed8709d341fc0e5d8 Mon Sep 17 00:00:00 2001 From: Roger Hunt Date: Sun, 21 Jun 2026 04:56:38 -0700 Subject: [PATCH] Submission: Humanized AI and Algorithmic Coordination in Healthcare Practice --- submissions/ii.md | 43 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 43 insertions(+) create mode 100644 submissions/ii.md diff --git a/submissions/ii.md b/submissions/ii.md new file mode 100644 index 0000000..6b8c1ec --- /dev/null +++ b/submissions/ii.md @@ -0,0 +1,43 @@ +# Humanized AI and Algorithmic Coordination in Healthcare Practice + +**Authors:** Phillip M. Rowell II, Chief Analytics Officer, Health Catalyst +**Contact:** phillip.rowell@willowgroveint.com +**Type:** Practitioner report +**Track:** TR.04 — Trust, Opacity & Governance +**Word count:** 372 +**Keywords:** algorithmacy, humanized AI, algorithmic accountability, healthcare analytics, human-in-the-loop, governance, platform mediation +**Conflicts of interest:** None + +--- + +## Abstract + +Algorithmic systems now quietly coordinate clinical and operational decisions across modern health systems, from triage prioritization to staffing, discharge planning, and resource allocation. In many organizations, these systems do more than support work: they function as algorithmic third parties, shaping choices, sequencing actions, and redistributing authority across clinicians, administrators, and technical infrastructures. Yet the models and platforms involved are often opaque to the people expected to rely on them. Recommendations appear with institutional force, while the reasoning, weighting, and design assumptions behind them remain difficult to inspect, question, or contest. The result is not simply a problem of technical complexity. It is a coordination problem and, more importantly, a governance problem. + +Drawing on direct experience in healthcare analytics leadership, this practitioner report examines how platform mediation alters decision-making in environments where trust, safety, and accountability are non-negotiable. When an algorithm influences who gets seen first, how staff are deployed, or where resources are directed, the practical question is not whether the system is advanced, but whether the organization can still explain how judgment is being exercised. As algorithmic outputs become embedded in routine workflows, responsibility can quietly diffuse: clinicians may defer to tools they cannot interrogate, operational leaders may implement systems they did not design, and technical teams may optimize for performance without adequate mechanisms for institutional oversight. + +The central claim of this report is that opacity should be understood not as an incidental feature of sophisticated systems but as a governance failure with organizational consequences. In response, I argue for a model of humanized AI grounded in algorithmacy and meaningful human-in-the-loop design. Humanized AI is not a branding exercise or a superficial call for empathy in technology. It is an operational commitment to preserving human authority, interpretability, and contestability inside algorithmically mediated work. Algorithmacy, in turn, names the practical capacity leaders and workers need in order to read, question, and govern the systems coordinating their decisions. + +This report contributes a practitioner perspective to debates on trust, opacity, and governance by showing how healthcare makes these issues visible in especially high-stakes form. Its conclusion is practical: organizations should treat explainability, escalation pathways, and accountable oversight not as optional design extras, but as core requirements for any system permitted to coordinate consequential work. + +## Outline + +Algorithmic systems increasingly coordinate clinical and operational work in healthcare, often functioning as quiet third parties in triage, staffing, and resource allocation. +Platform mediation reshapes how clinicians and administrators make decisions, narrowing visibility into how recommendations are formed and acted upon. +Accountability erodes when system outputs cannot be traced, questioned, or meaningfully interpreted by the people expected to rely on them. +This submission reframes opacity not as a mere technical problem but as a governance failure with direct consequences for trust, safety, and organizational responsibility. +It advances a practitioner case for humanized AI grounded in algorithmacy: the capacity to read, challenge, and govern algorithmic systems in real settings. +It argues that human-in-the-loop design must restore meaningful human authority rather than serve as symbolic oversight after the fact. +It concludes with practical governance implications for health leaders who need to embed transparency, explainability, and accountability into everyday operations now. + +## Author bios + +Phillip M. Rowell II is Chief Analytics Officer at Health Catalyst, where he focuses on algorithmacy, humanized AI, and the governance of algorithmic systems in healthcare. His work examines how opaque platform mediation erodes accountability in clinical and operational decision-making—and how human-in-the-loop design can restore it + +## Statement on review policy + +By submitting, I/we acknowledge that this submission and all reviews of it will be public on this repository under the conference's open-review policy (see [README.md](../README.md#review-policy)). + +--- + +*Submitted via the web form at algorithmacy.org on behalf of the listed authors; the pull request was opened by the conference account. The PR timestamp is the priority record.*