ksb-d08-k0045

name	KSB-D08-K0045
description	Human-in-the-Loop (HITL) Requirements: Comprehensive understanding of human-in-the-loop requirements for AI-driven pharmacovigilance includ...
version	1.0.0
domain	D08
domain_name	Signal Detection & Management
type	Knowledge
proficiency_level	L3
bloom_level	apply
triggers	["explain human-in-the-loop (hitl) requirements","what is human-in-the-loop (hitl) requirements","apply human-in-the-loop (hitl) requirements"]
epa_mapping	EPA-04, EPA-05
cpa_mapping	CPA-02
regulatory_refs

KSB-D08-K0045: Human-in-the-Loop (HITL) Requirements

Overview

Domain: D08 - Signal Detection & Management Type: Knowledge Proficiency Level: L3 (Competent - Minimal supervision) Bloom Level: Apply

Description

Comprehensive understanding of human-in-the-loop requirements for AI-driven pharmacovigilance including CIOMS WG XIV oversight principles, levels of automation requiring human review, critical decision points mandating human judgment, and regulatory expectations for maintaining human control over safety-critical AI outputs. Includes knowledge of HITL design patterns: human-on-the-loop (monitoring), human-in-the-loop (approval), and human-in-command (override) configurations.

Context

• Major Section: AI Integration and Advanced Analytics
• Section: Human Oversight in AI-Driven Signal Detection

EPA Mapping

• EPA-04:3009-3010
• EPA-05:3011-3012

CPA Pathway

• CPA-02

Regulatory References

• General PV competency

Instructions

When this skill is activated, Claude should:

1. Demonstrate L3 proficiency in human-in-the-loop (hitl) requirements
2. Apply apply level cognitive skills to apply the topic
3. Reference relevant regulatory guidance (general PV standards)
4. Connect to related EPAs: EPA-04, EPA-05

Key Competencies

• Comprehensive understanding of human-in-the-loop requirements for AI-driven pharmacovigilance including CIOMS WG XIV oversight principles, levels of automation requiring human review, critical decision points mandating human judgment, and regulatory expectations for maintaining human control over safety-critical AI outputs. Includes knowledge of HITL design patterns: human-on-the-loop (monitoring), human-in-the-loop (approval), and human-in-command (override) configurations.

Assessment Criteria

• Can apply core concepts independently
• Demonstrates understanding of regulatory context
• Applies knowledge appropriately to PV scenarios

Related Skills

• Other D08 skills in Human Oversight in AI-Driven Signal Detection
• Cross-domain integrations per DAG architecture

---

Generated from PV KSB Framework v1.0 | 2025-12-31