Skip to content

AI Interaction Design

LMU/COICP Example

AI Interaction Design Example

Define LMU/COICP AI Incident Summary as deferred/disabled for RC-001 while preserving future AI draft-summary rules, input limits, output handling, human review behavior, failure behavior, and prohibitions.

ES-105 AI Interaction Deferred / Disabled

Project

LMU Campus Operations and Incident Coordination Platform

Document control

Field Value
Artifact owner AI reviewer
Primary reviewers COICP Product Engineer, Architecture Review Board chair, IT security reviewer, Compliance reviewer
Status Accepted with AI Incident Summary deferred/disabled
Last updated 2026-07-06
Related Engineering Stage ES-105 — Design
Project workspace target docs/project-workspace/design/ai_interaction_design.md

Design context

Field Value
Source architecture package ES-104 — Architecture
Design baseline produced LMU-COICP-DES-001
Implementation baseline expected next LMU-COICP-IMPL-001
Integrated baseline expected later LMU-COICP-INT-001
Release candidate expected later LMU-COICP-RC-001
Downstream release posture Internal engineering review only
Planned downstream review environment LMU-COICP Internal Engineering Review Environment
Planned downstream dataset Spring Semester Synthetic Incident Dataset
Planned downstream synthetic records 421
Planned downstream reviewer accounts 24
Planned downstream review window March 18–22, 2026
AI Incident Summary Deferred / disabled
Operational pilot Not approved

Current AI behavior

AI Incident Summary is disabled for LMU-COICP-DES-001, LMU-COICP-IMPL-001, LMU-COICP-INT-001, and LMU-COICP-RC-001.

AI disabled user experience

Reviewer Action Scenario Expected Result
Campus Operations reviewer looks for summary button COICP-SYN-118 no AI summary action displayed
AI reviewer checks feature state RC-001 configuration AI Incident Summary disabled
ARB reviews evidence timeline COICP-SYN-118 no AI-generated official event present
Compliance reviewer inspects data flow Spring Semester Synthetic Incident Dataset no active AI data path

Future AI draft summary design

If a future readiness cycle enables AI, the design must support authorized request, approved incident context, marked draft generation, human accept/edit/reject, recorded human action, and no official record use without human acceptance. For COICP-SYN-118, future AI may summarize creation, Facilities handoff, status updates, and closure rationale; it may not create official closure, assign responsibility, determine blame, infer discipline, or send communications.

AI prohibitions

  • no AI-generated official summary in RC-001;
  • no hidden AI content;
  • no AI-generated closure, owner assignment, escalation, emergency guidance, or communications approval;
  • no AI use with prohibited or real data;
  • no AI activation without future release, operations, deployment, monitoring, stewardship, and governance evidence.

Continue to Error and Exception Design

Define safe responses to realistic LMU failure cases.

Open Error and Exception Design →