AI Human Oversight Standard

Document Type: Standard
Standard ID: STD-AI-007
Standard Title: AI Human Oversight Standard
Version: 1.0
Effective Date: 2025-08-01
Next Review Date: 2026-08-01
Review Frequency: Annually or upon regulatory change
Parent Policy: POL-AI-001 - Artificial Intelligence Policy
Owner: AI Risk Manager
Approved By: AI Governance Committee Chair
Status: Draft
Classification: Internal Use Only

---

TABLE OF CONTENTS

1. Document History
2. Objective
3. Scope and Applicability
4. Control Standard
5. Supporting Procedures
6. Compliance
7. Roles and Responsibilities
8. Exceptions
9. Enforcement
10. Key Performance Indicators (KPIs)
11. Training Requirements
12. Definitions
13. Link with AI Act and ISO42001

---

DOCUMENT HISTORY

Version	Date	Author	Changes	Approval Date	Approved By
0.1	2025-07-02	AI Risk Manager	Initial draft	-	-
0.2	2025-07-18	AI Risk Manager	Added oversight measures details	-	-
0.3	2025-07-30	AI Risk Manager	Incorporated stakeholder feedback	-	-
1.0	2025-08-01	AI Risk Manager	Final version approved - GRC restructured	2025-07-25	Jane Doe, AI Governance Committee Chair

---

OBJECTIVE

This standard defines requirements for designing AI systems to enable effective human oversight and intervention in compliance with EU AI Act Article 14.

Primary Goals:

• Design AI systems with effective human oversight per Article 14(3) and 14(4)
• Ensure oversight personnel have necessary capabilities per Article 14(4)
• Implement and maintain effective human oversight throughout AI system lifecycle
• Monitor oversight effectiveness and continuously improve

---

SCOPE AND APPLICABILITY

2.1 Mandatory Applicability

This standard is mandatory for:

• All high-risk AI systems (EU AI Act Article 14)
• All AI systems with significant impact on health, safety, or fundamental rights

2.2 Recommended Applicability

This standard is recommended for:

• All AI systems for best practices
• Limited-risk and minimal-risk AI systems (voluntary human oversight)

2.3 Oversight Requirements Covered

• Human oversight design and measures
• Oversight personnel capabilities
• Oversight implementation and monitoring
• Oversight effectiveness assessment

2.4 Out of Scope

• General system monitoring (covered by operations standards)
• Non-AI system oversight (covered by other oversight standards)
• Oversight outside EU AI Act scope

---

CONTROL STANDARD

Control HO-001: Human Oversight Measures Design

Control ID: HO-001
Control Name: Human Oversight Measures Implementation
Control Type: Preventive
Control Frequency: Per high-risk AI system, during design phase
Risk Level: High

Control Objective

Design AI systems with effective human oversight measures per Article 14(3) and 14(4) to enable humans to understand, interpret, override, and intervene in AI system operations, ensuring human control and preventing automation bias.

Control Requirements

CR-001.1: Mandatory Oversight Measures

Implement all required oversight measures per Article 14(3) and 14(4).

Required Measures (Article 14(3) and 14(4)):

Measure	Description	Implementation	Verification
Understand Capabilities	Fully understand AI system capabilities and limitations	Training, documentation, interface design	Knowledge assessment
Awareness of Automation Bias	Remain aware of automation bias tendency	Training, reminders, alerts	Awareness verification
Interpret Outputs	Correctly interpret AI system outputs	Clear output presentation, explanations	Interpretation testing
Decision Not to Use	Decide not to use AI system in particular situation	Manual override option, decision support	Override capability testing
Override Decision	Override or reverse AI system decision	Override mechanism, confirmation	Override functionality testing
Intervene in Operation	Intervene in AI system operation	Intervention controls, pause capability	Intervention testing
Stop Operation	Stop AI system operation	Stop button, emergency stop	Stop functionality testing

Mandatory Actions:

• Design oversight interface
• Implement override capability
• Enable intervention mechanisms
• Provide stop functionality
• Test oversight effectiveness
• Document oversight design
• Obtain approval before deployment

Oversight Interface Requirements:

• Clear display of AI system outputs
• Confidence scores and explanations
• Override controls (prominent, accessible)
• Intervention controls
• Stop button (prominent, accessible)
• Alert mechanisms for critical situations
• Decision support information

Evidence Required:

• Human oversight design document
• Interface specifications
• Override mechanism documentation
• Intervention mechanism documentation
• Stop functionality documentation
• Test results
• Approval records

Audit Verification:

• Verify all oversight measures implemented
• Confirm oversight interface designed
• Check override capability functional
• Validate intervention mechanisms operational
• Verify stop functionality tested
• Check approval obtained before deployment

---

Control HO-002: Oversight Personnel Capabilities

Control ID: HO-002
Control Name: Oversight Personnel Competence and Authority
Control Type: Preventive
Control Frequency: Per oversight assignment, ongoing
Risk Level: High

Control Objective

Ensure oversight personnel have necessary competence, training, authority, and resources per Article 14(4) to effectively perform oversight functions and make informed decisions about AI system use.

Control Requirements

CR-002.1: Personnel Qualification and Authority

Assign qualified personnel with necessary competence, training, and authority.

Required Capabilities (Article 14(4)):

Capability	Requirement	Implementation	Verification
Competence	Necessary competence for oversight role	Qualification assessment, certification	Competency validation
Training	Appropriate training on AI system and oversight	Initial and ongoing training	Training completion records
Authority	Authority to override, intervene, stop	Authority delegation, role assignment	Authority verification
Resources	Adequate resources to perform oversight	Time allocation, tools, support	Resource allocation verification
Technical Measures	Technical measures enabling oversight	Oversight tools, interfaces	Technical capability verification

Mandatory Actions:

• Identify oversight personnel requirements
• Assess personnel qualifications
• Assign qualified personnel
• Provide initial training
• Grant necessary authority
• Allocate adequate resources
• Implement technical measures
• Verify capabilities

Oversight Personnel Requirements:

Role	Competence Requirements	Training Requirements	Authority Level
Primary Oversight	Deep AI system knowledge, decision-making	Comprehensive training (16 hours)	Full override, intervene, stop
Secondary Oversight	AI system knowledge, monitoring	Standard training (8 hours)	Override, intervene (with approval)
Emergency Oversight	Basic AI system knowledge, emergency response	Emergency training (4 hours)	Emergency stop only

Evidence Required:

• Oversight personnel assignments
• Qualification assessment records
• Training records
• Authority delegation records
• Resource allocation records
• Technical implementation records
• Capability verification records

Audit Verification:

• Verify oversight personnel assigned
• Confirm qualifications assessed
• Check training completed
• Validate authority granted
• Verify resources allocated
• Check technical measures implemented

---

Control HO-003: Oversight Implementation and Operations

Control ID: HO-003
Control Name: Human Oversight Operations and Monitoring
Control Type: Detective
Control Frequency: Continuous, ongoing
Risk Level: Medium

Control Objective

Implement and maintain effective human oversight operations throughout AI system lifecycle, ensuring oversight personnel are available, trained, and performing oversight functions effectively.

Control Requirements

CR-003.1: Ongoing Oversight Operations

Maintain effective oversight operations with qualified personnel.

Operational Requirements:

Requirement	Specification	Implementation
Personnel Availability	Oversight personnel available during AI system operation	Staffing schedules, on-call coverage
Ongoing Training	Regular training updates and refreshers	Annual training, update sessions
Performance Monitoring	Monitor oversight personnel performance	Performance reviews, metrics
Role Clarity	Clear roles and responsibilities	Job descriptions, RACI matrices
Adequate Staffing	Sufficient personnel for oversight load	Staffing analysis, capacity planning

Mandatory Actions:

• Maintain oversight personnel assignments
• Provide ongoing training (annually minimum)
• Monitor performance regularly
• Review roles and responsibilities
• Ensure adequate staffing levels
• Conduct regular oversight reviews
• Document oversight activities

Oversight Activity Logging:

• Oversight personnel assignments
• Training completion
• Override actions
• Intervention actions
• Stop actions
• Performance metrics
• Review outcomes

Evidence Required:

• Personnel assignment records
• Ongoing training records
• Performance review records
• Oversight activity logs
• Staffing analysis records
• Review records

Audit Verification:

• Verify oversight personnel available
• Confirm ongoing training provided
• Check performance monitored
• Validate adequate staffing
• Verify oversight activities logged

---

Control HO-004: Oversight Effectiveness Monitoring

Control ID: HO-004
Control Name: Oversight Effectiveness Assessment and Improvement
Control Type: Detective
Control Frequency: Monthly monitoring, quarterly reviews
Risk Level: Medium

Control Objective

Monitor oversight effectiveness, track metrics, analyze patterns, and implement improvements to ensure human oversight remains effective and achieves its objectives.

Control Requirements

CR-004.1: Effectiveness Monitoring and Analysis

Monitor and analyze oversight effectiveness continuously.

Monitoring Requirements:

Metric	Target	Measurement	Frequency
Override Rate	Monitor trends	% of decisions overridden	Monthly
Intervention Rate	Monitor trends	% of operations intervened	Monthly
Stop Rate	Monitor trends	% of operations stopped	Monthly
Response Time	< 5 minutes	Average time to override/intervene	Monthly
Oversight Coverage	100%	% of AI operations with oversight	Daily
Training Completion	100%	% of personnel trained	Quarterly
Effectiveness Score	≥ 95%	Composite effectiveness score	Quarterly

Mandatory Actions:

• Track oversight activities
• Calculate effectiveness metrics
• Analyze trends and patterns
• Identify improvement opportunities
• Implement improvements
• Report to management
• Conduct quarterly effectiveness reviews

Effectiveness Analysis:

Analysis Type	Purpose	Frequency
Override Pattern Analysis	Understand when/why overrides occur	Monthly
Intervention Analysis	Understand intervention patterns	Monthly
Response Time Analysis	Identify response time issues	Monthly
Training Effectiveness	Assess training impact	Quarterly
Overall Effectiveness	Comprehensive effectiveness assessment	Quarterly

Evidence Required:

• Oversight activity logs
• Metrics dashboard
• Analysis reports (monthly)
• Effectiveness reports (quarterly)
• Improvement initiatives
• Management reports

Audit Verification:

• Verify oversight activities tracked
• Confirm metrics calculated
• Check analysis conducted
• Validate improvements implemented
• Verify reports generated

---

SUPPORTING PROCEDURES

This standard is implemented through the following detailed procedures:

Procedure PROC-AI-HO-001: Human Oversight Design Procedure

Purpose: Define step-by-step process for designing human oversight
Owner: AI Risk Manager
Implements: Controls HO-001, HO-002

Procedure Steps:

1. Assess AI system risks and oversight needs
2. Design oversight measures - Control HO-001
3. Design oversight interface
4. Identify oversight personnel requirements - Control HO-002
5. Assign oversight personnel
6. Provide training
7. Grant authority
8. Test oversight effectiveness
9. Obtain approval

Outputs:

• Human oversight design document
• Oversight interface specifications
• Personnel assignments
• Training records
• Approval records

---

Procedure PROC-AI-HO-002: Oversight Effectiveness Review Procedure

Purpose: Define process for reviewing oversight effectiveness
Owner: AI Risk Manager
Implements: Controls HO-003, HO-004

Procedure Steps:

1. Collect oversight activity data - Control HO-003
2. Calculate effectiveness metrics - Control HO-004
3. Analyze trends and patterns
4. Identify improvement opportunities
5. Implement improvements
6. Report to management
7. Conduct quarterly reviews

Outputs:

• Effectiveness analysis reports
• Improvement initiatives
• Management reports

---

COMPLIANCE

5.1 Compliance Monitoring

Monitoring Approach: Continuous automated monitoring supplemented by monthly manual reviews and quarterly comprehensive audits.

Compliance Metrics:

Metric	Target	Measurement Method	Frequency	Owner
Oversight Measures Implementation	100%	% of high-risk AI with all oversight measures	Monthly	AI Risk Manager
Oversight Personnel Assignment	100%	% of AI systems with assigned oversight personnel	Monthly	AI Risk Manager
Oversight Personnel Training	100%	% of oversight personnel trained	Quarterly	AI Risk Manager
Oversight Coverage	100%	% of AI operations with oversight	Daily	AI Risk Manager
Oversight Effectiveness	≥95%	Composite effectiveness score	Quarterly	AI Risk Manager
Override Response Time	< 5 minutes	Average time to override	Monthly	AI Risk Manager
Intervention Success Rate	≥98%	% of successful interventions	Monthly	AI Risk Manager

Monitoring Tools:

• Oversight Activity Dashboard
• Effectiveness Metrics Dashboard
• Compliance Reports
• Monthly compliance reports
• Quarterly AI Governance Committee reviews

---

5.2 Internal Audit Requirements

Audit Frequency: Annually (minimum)

Audit Scope:

• Oversight measures implementation
• Oversight personnel qualifications
• Oversight operations effectiveness
• Oversight monitoring and analysis
• Controls effectiveness (HO-001 through HO-004)

Audit Activities:

• Review 100% of high-risk AI for oversight implementation
• Test oversight measures functionality
• Verify oversight personnel qualifications
• Review oversight activity logs
• Test override/intervention capabilities
• Analyze effectiveness metrics
• Interview oversight personnel
• Test oversight response times

Audit Outputs:

• Annual Human Oversight Audit Report
• Findings and recommendations
• Corrective action plans for deficiencies

---

5.3 External Audit / Regulatory Inspection

Preparation:

• Maintain audit-ready oversight documentation at all times
• Designate AI Risk Manager and Legal as regulatory liaisons
• Prepare standard response procedures for authority requests

Provide to Auditors/Regulators:

• Human oversight design documents
• Oversight personnel assignments
• Training records
• Oversight activity logs
• Effectiveness reports
• Oversight procedures
• Internal audit reports
• Evidence of controls execution

Authority Request Response:

• Acknowledge request within 1 business day
• Provide requested documentation within 5 business days
• Coordinate through Legal and AI Risk Manager
• Document all interactions with authorities

---

ROLES AND RESPONSIBILITIES

6.1 RACI Matrix

Activity	AI Risk Manager	AI System Owner	Oversight Personnel	Product Director	HR	Training
Oversight Design	R/A	A	C	C	I	I
Oversight Measures Implementation	R	A	C	C	I	I
Personnel Assignment	R	A	I	I	R	I
Personnel Training	R	A	A	I	C	R
Authority Grant	R	A	I	I	C	I
Oversight Operations	R	A	R	I	I	I
Effectiveness Monitoring	R	A	C	I	I	I
Performance Reviews	R	A	A	I	R	I

RACI Legend:

• R = Responsible (does the work)
• A = Accountable (ultimately answerable)
• C = Consulted (provides input)
• I = Informed (kept up-to-date)

---

6.2 Role Descriptions

AI Risk Manager

• Primary Responsibility: Owns oversight framework, ensures compliance
• Key Activities:
  • Establishes oversight framework
  • Designs oversight measures
  • Monitors oversight effectiveness
  • Reports oversight metrics
• Required Competencies: EU AI Act Article 14, oversight design, risk management

AI System Owner

• Primary Responsibility: Accountable for oversight of their AI system
• Key Activities:
  • Ensures oversight implemented
  • Assigns oversight personnel
  • Monitors oversight operations
  • Participates in effectiveness reviews
• Required Competencies: AI system knowledge, oversight requirements

Oversight Personnel

• Primary Responsibility: Perform human oversight functions
• Key Activities:
  • Monitor AI system operations
  • Override AI decisions when needed
  • Intervene in AI operations
  • Stop AI operations when necessary
  • Document oversight activities
• Required Competencies: AI system knowledge, decision-making, oversight procedures

Product Director

• Primary Responsibility: Provides resources and support for oversight
• Key Activities:
  • Allocates resources for oversight
  • Supports oversight implementation
  • Reviews oversight effectiveness
• Required Competencies: Resource management, oversight awareness

HR

• Primary Responsibility: Manages oversight personnel assignments
• Key Activities:
  • Identifies qualified personnel
  • Manages assignments
  • Supports performance reviews
• Required Competencies: Personnel management, qualification assessment

Training

• Primary Responsibility: Provides oversight training
• Key Activities:
  • Develops training programs
  • Delivers training
  • Assesses training effectiveness
• Required Competencies: Training development, AI system knowledge

---

EXCEPTIONS

7.1 Exception Philosophy

Human oversight is a critical regulatory compliance activity for high-risk AI systems. Exceptions are granted restrictively and only where compensating controls adequately mitigate risks.

---

7.2 Allowed Exceptions

The following exceptions may be granted with proper justification and approval:

Exception Type	Justification Required	Maximum Duration	Approval Authority	Compensating Controls
Reduced Oversight (Minimal-Risk AI)	AI system clearly minimal-risk; reduced oversight sufficient	Permanent	AI Risk Manager	Document rationale; Annual re-confirmation
Extended Training Timeline	Resource constraints prevent timely training	30 days	AI Risk Manager + HR	Interim oversight; Accelerated training plan

---

7.3 Prohibited Exceptions

The following exceptions cannot be granted under any circumstances:

❌ Skipping oversight for high-risk AI - Mandatory per Article 14, no exceptions
❌ Disabling override capability - Required for human control
❌ Operating without oversight personnel - Required for effective oversight
❌ Skipping oversight training - Required for personnel competence
❌ Removing stop functionality - Required for safety

---

7.4 Exception Request Process

Step 1: Submit Exception Request

• Complete Exception Request Form (FORM-AI-EXCEPTION-001)
• Include business justification
• Propose compensating controls
• Specify duration requested
• Attach risk assessment

Step 2: Risk Assessment

• AI Risk Manager assesses risk of granting exception
• Evaluates adequacy of compensating controls
• Documents residual risk

Step 3: Approval

• Route to appropriate approval authority based on exception type
• AI Risk Manager approval: Minor exceptions
• AI Risk Manager + AI Governance Committee: Significant exceptions
• AI Governance Committee: Critical exceptions

Step 4: Documentation and Monitoring

• Document exception in Exception Register
• Assign exception owner
• Set review date
• Monitor compensating controls
• Report exceptions quarterly to AI Governance Committee

Step 5: Exception Review and Closure

• Review exception at specified review date
• Assess if exception still needed
• Close exception when normal oversight completed
• Document lessons learned

---

ENFORCEMENT

8.1 Non-Compliance Consequences

Violation	Severity	Consequence	Remediation Required
High-risk AI without oversight	Critical	Immediate suspension until oversight implemented	Implement oversight within 5 business days; Root cause analysis
Missing override capability	Critical	Immediate correction; Compliance gap assessment	Implement override within 5 business days
Oversight personnel not trained	High	Escalation to AI Governance Committee; Freeze on operations	Complete training within 10 business days
Oversight not operational	High	Escalation to management	Restore oversight within 24 hours
Missing stop functionality	Critical	Immediate correction	Implement stop functionality within 5 business days
Oversight effectiveness below target	Medium	Written warning; Improvement plan	Implement improvements within 30 business days

---

8.2 Escalation Procedures

Level 1: AI Risk Manager

• Minor procedural violations
• Training delays
• Timeline delays < 5 days
• Action: Written warning, corrective action required

Level 2: AI Risk Manager + AI Governance Committee

• Repeated violations
• Missing oversight measures
• Untrained personnel
• Action: Formal review, corrective action plan, management notification

Level 3: AI Governance Committee

• High-risk AI without oversight
• Missing override/stop capability
• Critical compliance failures
• Action: Immediate AI system suspension, investigation, disciplinary action

Level 4: Executive Management + Legal

• Potential regulatory enforcement action
• Significant legal liability
• Reputational risk
• Action: Executive crisis management, legal strategy, regulatory engagement

---

8.3 Immediate Escalation Triggers

Escalate immediately to AI Governance Committee + Legal if:

• ⚠️ High-risk AI system operating without oversight
• ⚠️ Override or stop capability disabled
• ⚠️ Oversight personnel unable to perform functions
• ⚠️ Regulatory inquiry or inspection related to oversight
• ⚠️ Critical oversight failure leading to incident

---

8.4 Disciplinary Actions

Individuals responsible for oversight violations may be subject to:

• Verbal or written warning
• Mandatory retraining
• Performance improvement plan
• Reassignment of responsibilities
• Suspension (with pay during investigation)
• Termination (for egregious violations, e.g., knowingly disabling oversight)

Factors Considered:

• Intent (knowing violation vs. honest mistake)
• Severity of violation
• Impact (actual or potential)
• Cooperation with remediation
• Prior violation history

---

KEY PERFORMANCE INDICATORS (KPIs)

9.1 Human Oversight KPIs

KPI ID	KPI Name	Definition	Target	Measurement Method	Frequency	Owner	Reporting To
KPI-HO-001	Oversight Measures Implementation	% of high-risk AI with all oversight measures	100%	(# AI with all measures / # high-risk AI) × 100	Monthly	AI Risk Manager	AI Governance Committee
KPI-HO-002	Oversight Personnel Assignment	% of AI systems with assigned oversight personnel	100%	(# AI with personnel / # total AI) × 100	Monthly	AI Risk Manager	Management
KPI-HO-003	Oversight Personnel Training	% of oversight personnel trained	100%	(# trained personnel / # total personnel) × 100	Quarterly	AI Risk Manager	AI Governance Committee
KPI-HO-004	Oversight Coverage	% of AI operations with oversight	100%	(# operations with oversight / # total operations) × 100	Daily	AI Risk Manager	Management
KPI-HO-005	Oversight Effectiveness	Composite effectiveness score	≥95%	Weighted average of effectiveness metrics	Quarterly	AI Risk Manager	AI Governance Committee
KPI-HO-006	Override Response Time	Average time to override AI decision	< 5 minutes	Σ (override time) / # overrides	Monthly	AI Risk Manager	Management
KPI-HO-007	Intervention Success Rate	% of successful interventions	≥98%	(# successful interventions / # total interventions) × 100	Monthly	AI Risk Manager	Management
KPI-HO-008	Stop Response Time	Average time to stop AI system	< 2 minutes	Σ (stop time) / # stops	Monthly	AI Risk Manager	Management
KPI-HO-009	Oversight Activity Logging	% of oversight activities logged	100%	(# logged activities / # total activities) × 100	Monthly	AI Risk Manager	Management
KPI-HO-010	Training Effectiveness	Training effectiveness score	≥90%	Post-training assessment scores	Quarterly	Training	AI Risk Manager

---

9.2 KPI Dashboards and Reporting

Real-Time Dashboard (AI Risk Manager access)

• Current oversight coverage
• Oversight personnel availability
• Override/intervention rates
• Response times
• System health

Monthly Management Report

• KPI-HO-001, 002, 004, 006, 007, 008, 009
• Trend analysis (vs. previous month)
• Issues and risks
• Planned actions

Quarterly AI Governance Committee Report

• All KPIs
• Oversight effectiveness assessment
• Training completion status
• Internal audit findings (if conducted)
• Exception register review

Annual Executive Report

• Full-year KPI performance
• Oversight maturity assessment
• Strategic recommendations
• Regulatory outlook

---

9.3 KPI Thresholds and Alerts

KPI	Green (Good)	Yellow (Warning)	Red (Critical)	Alert Action
Oversight Measures Implementation	100%	95-99%	< 95%	Red: Immediate escalation to AI Governance Committee Chair
Oversight Coverage	100%	95-99%	< 95%	Red: Immediate escalation to AI Governance Committee
Oversight Effectiveness	≥95%	90-94%	< 90%	Yellow: Improvement plan; Red: Escalate to AI Governance Committee
Override Response Time	< 5 minutes	5-10 minutes	> 10 minutes	Red: Escalate to AI Governance Committee

---

TRAINING REQUIREMENTS

10.1 Training Program Overview

All personnel involved in human oversight must complete role-specific training to ensure competency in EU AI Act Article 14 requirements, oversight procedures, and oversight tools.

---

10.2 Role-Based Training Requirements

Role	Training Course	Duration	Content	Frequency	Assessment Required
Oversight Personnel (Primary)	Comprehensive Oversight Training	16 hours	EU AI Act Article 14; AI system knowledge; Override procedures; Intervention procedures; Decision-making; Automation bias	Initial + annually	Yes - Written exam (≥90%) + Practical exercise
Oversight Personnel (Secondary)	Standard Oversight Training	8 hours	Oversight basics; Override procedures; Monitoring	Initial + annually	Yes - Knowledge check (≥80%) + Practical exercise
Oversight Personnel (Emergency)	Emergency Oversight Training	4 hours	Emergency stop procedures; Basic oversight	Initial + annually	Yes - Practical exercise
AI Risk Manager	Oversight Management Training	12 hours	Oversight framework; Effectiveness monitoring; Personnel management	Initial + annually	Yes - Written exam (≥90%)
AI System Owners	Oversight Overview	4 hours	Oversight requirements; Responsibilities; Personnel assignment	At onboarding + annually	Yes - Knowledge check (≥80%)

---

10.3 Training Content by Topic

EU AI Act Article 14 Requirements

• Oversight measures framework (Article 14(3))
• Oversight capabilities (Article 14(4))
• Compliance obligations

Oversight Measures

• Understanding AI capabilities and limitations
• Automation bias awareness
• Interpreting AI outputs
• Override procedures
• Intervention procedures
• Stop procedures

Decision-Making

• When to override
• When to intervene
• When to stop
• Decision support
• Risk assessment

Oversight Tools

• Oversight interface
• Override controls
• Intervention controls
• Monitoring tools
• Logging tools

---

10.4 Training Delivery Methods

Initial Training:

• Instructor-led classroom or virtual training
• Includes interactive exercises and case studies
• Hands-on practice with oversight tools
• Group discussions of complex scenarios

Annual Refresher:

• E-learning modules for core content review
• Live update sessions for regulatory changes
• Case study reviews of recent oversight activities
• Knowledge assessment

On-the-Job Training:

• Mentoring for new oversight personnel
• Job shadowing during oversight operations
• Supervised oversight for first 10 AI operations

Just-in-Time Training:

• Quick reference guides and job aids
• Video tutorials on specific topics
• Help desk support from experienced oversight personnel

---

10.5 Training Effectiveness Measurement

Assessment Methods:

• Written exams for knowledge retention
• Practical exercises for skill application
• On-the-job observations for competency validation
• Feedback surveys for training quality

Competency Validation:

• Primary Oversight Personnel: Must demonstrate ability to override, intervene, and stop AI system with 100% success before independent oversight
• All personnel: Must pass knowledge assessments with minimum required scores

Training Metrics:

Metric	Target	Frequency
Training completion rate	100%	Quarterly
Assessment pass rate (first attempt)	≥ 90%	Per training
Training effectiveness score (survey)	≥ 4.0/5.0	Per training
Time to competency (Primary Oversight)	< 30 days	Per person

---

10.6 Training Records

Records Maintained:

• Training attendance records
• Assessment scores
• Competency validations
• Refresher training completion
• Individual training transcripts

Retention: 10 years (to align with EU AI Act documentation retention)

Access: HR, AI Risk Manager, Internal Audit, Competent Authorities (upon request)

---

DEFINITIONS

Term	Definition	Source
Human Oversight	Human ability to understand, monitor, override, and intervene in AI system operations	EU AI Act Article 14
Oversight Measures	Specific measures enabling human oversight (Article 14(3) and 14(4))	EU AI Act Article 14(3) and 14(4)
Oversight Capabilities	Required capabilities for oversight personnel (Article 14(4))	EU AI Act Article 14(4)
Override	Human action to reverse or change AI system decision	This Standard
Intervention	Human action to modify AI system operation	This Standard
Stop	Human action to halt AI system operation	This Standard
Automation Bias	Tendency to over-rely on automated systems and accept their outputs without question	This Standard
Oversight Personnel	Natural persons assigned to perform human oversight functions	This Standard
Oversight Effectiveness	Measure of how well human oversight achieves its objectives	This Standard

---

LINK WITH AI ACT AND ISO42001

12.1 EU AI Act Regulatory Mapping

This standard implements the following EU AI Act requirements:

EU AI Act Provision	Article	Requirement Summary	Implemented By (Controls)
Human Oversight	Article 14	Human oversight for high-risk AI	All controls (HO-001 through HO-004)
Oversight Measures	Article 14(3) and 14(4)	Specific measures enabling oversight	HO-001
Oversight Capabilities	Article 14(4)	Required capabilities for oversight personnel	HO-002
Oversight Implementation	Article 14	Effective oversight implementation	HO-003, HO-004

---

12.2 ISO/IEC 42001:2023 Alignment

This standard aligns with ISO/IEC 42001:2023 as follows:

ISO 42001 Clause	Requirement	Implementation in This Standard
Clause 6.1.2: AI system impact assessment	Assess human oversight impacts	HO-001, HO-002
Clause 7.2: Competence	Ensure personnel have appropriate competence	HO-002
Clause 8.2: AI system risk assessment	Human oversight in risk management	HO-001, HO-003
Clause 9.1: Monitoring, measurement, analysis, and evaluation	Monitor oversight effectiveness	HO-004

---

12.3 Relationship to Other Standards

This human oversight standard integrates with other AI Act standards:

Related Standard	Integration Point	Rationale
STD-AI-001: Classification	Classification determines if oversight required	High-risk AI requires Article 14 oversight
STD-AI-002: Risk Management	Oversight measures in risk management	Risk management identifies oversight needs
STD-AI-004: Technical Documentation	Oversight measures documented in Annex IV	Oversight information in technical documentation
STD-AI-006: Transparency	Oversight measures in Instructions for Use	Oversight information in Instructions for Use

---

12.4 References and Related Documents

EU AI Act (Regulation (EU) 2024/1689):

• Article 14: Human Oversight
• Article 14(2): Oversight Aim/Purpose (prevent or minimise risks)
• Article 14(3): Oversight Measures Framework (types of measures, commensurate with risks)
• Article 14(4): Oversight Capabilities (specific capability list for oversight personnel)

ISO/IEC Standards:

• ISO/IEC 42001:2023: Information technology — Artificial intelligence — Management system
• ISO/IEC 23894:2023: Information technology — Artificial intelligence — Guidance on risk management

Internal Documents:

• POL-AI-001: Artificial Intelligence Policy (parent policy)
• STD-AI-001: AI System Classification Standard
• STD-AI-002: AI Risk Management Standard
• STD-AI-004: AI Technical Documentation Standard
• STD-AI-006: AI Transparency Standard
• PROC-AI-HO-001, -002: Human oversight procedures

---

APPROVAL AND AUTHORIZATION

Role	Name	Title	Signature	Date
Prepared By	AI Risk Manager	AI Risk Manager	_________________	________
Reviewed By	Sarah Johnson	AI Act Program Manager	_________________	________
Reviewed By	Jane Doe	Chief Strategy & Risk Officer	_________________	________
Approved By	Jane Doe	AI Governance Committee Chair	_________________	________

Effective Date: 2025-08-01
Next Review Date: 2026-08-01
Review Frequency: Annually or upon regulatory change

---

END OF STANDARD STD-AI-007

---

This standard is a living document. Feedback and improvement suggestions should be directed to the AI Risk Manager.