Safety, Ergonomics, Risk, and Resilience

Introduction

Material-handling systems connect people, equipment, information, products, and facilities. They also concentrate operational risk. A poorly designed picking station can produce musculoskeletal disorders; an unsuitable pallet can collapse; an incompatible chemical package can leak; and a single unavailable conveyor or supplier can interrupt an entire distribution network. Safety, therefore, is not merely a compliance activity. It is a strategic design discipline that protects people, product integrity, service continuity, legal standing, and organizational reputation.

This chapter examines how supply-chain professionals can design material-handling and packaging systems that reduce injury, equipment incidents, hazardous-material exposure, product loss, and disruption. It integrates ergonomics, risk analysis, hierarchy-of-controls thinking, safety management, dangerous-goods governance, business continuity, and resilience engineering.

Learning objectives

By the end of this chapter, students should be able to:

1. Explain the relationship between safety, ergonomics, risk, packaging, and supply-chain resilience.
2. Diagnose common manual-handling and equipment hazards.
3. Apply ergonomic assessment concepts and interpret the NIOSH Revised Lifting Equation within its limits.
4. Conduct a structured risk assessment using severity, likelihood, exposure, and control effectiveness.
5. Select controls according to the hierarchy of controls and verify their effectiveness.
6. Explain major regulatory and governance responsibilities for hazardous materials, machinery, and warehouse operations.
7. Design continuity and recovery measures for material-handling and packaging systems.
8. Evaluate technology, human factors, and ethical issues in resilient safety design.

1. Safety as a supply-chain design objective

Traditional safety programs often focus on worker behavior: training, supervision, and personal protective equipment. These remain important, but a graduate-level approach recognizes that unsafe behavior is frequently an adaptation to system conditions. Workers may twist because storage locations are poorly positioned, bypass a guard because production targets are unrealistic, or overfill a carton because packaging specifications do not reflect actual product variation.

A useful systems view treats safety performance as an interaction among:

• People: capabilities, fatigue, training, communication, and decision-making.
• Tasks: force, posture, repetition, duration, pace, and complexity.
• Technology: conveyors, forklifts, robots, software, sensors, and controls.
• Environment: lighting, temperature, noise, floor condition, congestion, and layout.
• Materials and packaging: mass, dimensions, stability, sharp edges, chemicals, and labeling.
• Management systems: staffing, maintenance, purchasing, incentives, investigation, and learning.

Safety and productivity are not inherently opposites. Ergonomic improvements may reduce fatigue and errors; reliable preventive maintenance may increase equipment availability; stronger packaging may reduce both product damage and emergency cleanup. The relevant question is not whether safety costs money, but whether the organization is managing the total cost of risk.

Important terminology

A hazard is a source or situation with the potential to cause harm. Risk is the combination of the likelihood of harm and the severity of that harm. An incident is an unplanned event that may or may not result in injury or loss. A near miss is an event that caused no injury or damage but had the potential to do so. Resilience is the capability of a system to anticipate, withstand, absorb, adapt to, and recover from disruption while maintaining essential function.

A distinction is necessary between compliance and risk control. Compliance means meeting applicable legal or standard requirements. Risk control is broader: a compliant system can still be poorly designed if its hazards are not adequately understood, especially where regulations specify minimum rather than optimal performance.

2. Ergonomics and human-centered material handling

Ergonomics is the scientific discipline concerned with interactions among humans and other elements of a system, and the profession that applies theory, principles, data, and methods to optimize human well-being and overall system performance. In warehouses and packaging operations, ergonomics addresses physical, cognitive, and organizational demands.

2.1 Physical ergonomic risk factors

The principal physical risk factors are:

• Force: lifting, lowering, pushing, pulling, gripping, or resisting movement.
• Posture: bending, reaching, twisting, kneeling, squatting, or working overhead.
• Repetition: frequent performance of the same motion or cycle.
• Duration: length of exposure during a shift or across a work period.
• Vibration: hand-arm or whole-body vibration from tools, vehicles, or uneven floors.
• Contact stress: concentrated pressure from sharp edges, handles, or workstation surfaces.
• Environmental stress: heat, cold, poor lighting, noise, and restricted visibility.

Risk is often cumulative. A moderate force repeated thousands of times may be more harmful than a single high-force event. Conversely, a rare lift with extreme mass or an unstable load can create acute injury risk.

2.2 Design principles

A sound design places work within human capabilities rather than requiring workers to compensate for poor design. Practical principles include:

1. Store high-frequency items between approximately knee and shoulder height, while considering the specific worker population and task.
2. Bring the load close to the body and minimize horizontal reach.
3. Design handles and grasp points that support a neutral wrist and secure grip.
4. Reduce mass, package size, and required lifting frequency where feasible.
5. Use mechanical assistance for heavy, awkward, unstable, or repetitive loads.
6. Provide height-adjustable or appropriately dimensioned work surfaces.
7. Separate pedestrian and vehicle flows.
8. Use job rotation only as a supplement, not as a substitute for eliminating excessive exposure.
9. Design controls, displays, and alarms so that they are understandable under time pressure.

2.3 The NIOSH Revised Lifting Equation

For selected two-handed lifting tasks, the NIOSH Revised Lifting Equation estimates a recommended weight limit (RWL):

The lifting index is:

where:

The equation is intended primarily for smooth, two-handed lifting by healthy adult workers under defined conditions. It does not directly evaluate pushing, pulling, carrying, one-handed lifting, seated lifting, highly unstable loads, unpredictable loads, or many dynamic tasks. It also does not guarantee that a task is safe when LI ≤ 1; individual differences, fatigue, psychosocial conditions, and other hazards remain relevant. A higher lifting index indicates greater concern, and engineering redesign should normally be considered when the index exceeds 1.

The equation’s structure is multiplicative: poor performance in one factor can substantially reduce the recommended limit. This illustrates an important design principle—reducing reach distance may be more effective than simply telling workers to lift carefully.

Worked example: screening a lifting task

Suppose a carton weighs 12 kg and an assessment produces the following multipliers: HM = 0.625, VM = 0.93, DM = 0.85, AM = 0.86, FM = 0.95, and CM = 0.90. Using LC = 23 kg:

Therefore:

The task exceeds the reference screening value. Possible interventions include reducing carton mass, moving the load closer, improving coupling, reducing vertical travel, using a lift table, or redesigning the process. The result is a screening signal, not a diagnosis or a legal threshold.

3. Hazard identification and risk analysis

Risk assessment should be systematic, participative, and repeated when conditions change. A practical process is:

1. Define the task, population, operating state, and boundaries.
2. Identify normal, abnormal, maintenance, cleaning, startup, shutdown, and emergency conditions.
3. Identify hazards and credible initiating events.
4. Identify exposed people, assets, products, and environmental receptors.
5. Estimate risk using an explicit method.
6. Select controls according to the hierarchy of controls.
7. Assign ownership, deadlines, and verification criteria.
8. Reassess residual risk and monitor performance.

3.1 Risk matrices and their limitations

A common screening model is:

Severity and likelihood must be defined in advance. For example, a five-level scale may classify severity from first-aid injury to fatality and likelihood from rare to frequent. The score helps prioritize attention; it is not a measured probability and should not be interpreted as one unless the scale has been calibrated statistically.

A more complete assessment considers exposure and control reliability:

This expression is a managerial framework rather than a universal physical law. It is useful when a rare catastrophic event, frequent minor exposure, or fragile control system needs to be distinguished. Teams should document definitions rather than hide judgment inside numbers.

For failure-mode analysis, teams may use a risk-priority number:

RPN is useful for comparing failure modes within the same assessment, but it has known weaknesses. Multiplication can allow different combinations to produce the same score, and ordinal ratings do not possess equal numerical intervals. High-severity failures should therefore receive attention even when their RPN is not the largest.

3.2 Bow-tie thinking

Bow-tie analysis places a hazardous event at the center. Preventive barriers lie on the left between causes and the event; mitigative barriers lie on the right between the event and consequences. For example, in a forklift-pedestrian collision scenario:

• Causes may include poor visibility, mixed traffic, excessive speed, and inadequate training.
• Preventive barriers may include physical segregation, speed control, route design, warning systems, and authorization.
• Consequences may include injury, product damage, fire, or operational closure.
• Mitigative barriers may include impact protection, emergency response, medical arrangements, investigation, and recovery plans.

The value of bow-tie thinking is that it asks whether each critical barrier is present, independent, maintained, and effective—not merely whether a policy exists.

4. Hierarchy of controls and control verification

The hierarchy of controls ranks interventions from generally more reliable to less reliable:

1. Elimination: remove the hazardous manual lift or hazardous chemical from the process.
2. Substitution: use a less hazardous material, package, process, or technology.
3. Engineering controls: isolate people from hazards through guards, barriers, interlocks, lift devices, ventilation, or layout.
4. Administrative controls: procedures, training, scheduling, inspections, signage, and supervision.
5. Personal protective equipment: gloves, eye protection, footwear, respiratory protection, or fall protection.

PPE is essential for residual risk but is the least reliable as a primary control because it depends on correct selection, fit, availability, maintenance, and consistent human use. A mature design uses several independent layers where consequences are severe.

Control verification should specify what success looks like. Examples include measured force below a design criterion, a guard that prevents access during operation, a chemical package passing required performance tests, or a pedestrian route remaining unobstructed during peak volume. Audits should test actual conditions, not only documentation.

5. Equipment, vehicle, and packaging safety

5.1 Material-handling equipment

Common equipment hazards include struck-by events, caught-in points, tip-over, falling loads, unexpected startup, electrical energy, and inadequate maintenance. Controls include:

• equipment selection based on load, aisle, floor, gradient, environment, and operator visibility;
• machine guarding and interlocks;
• lockout/tagout or an equivalent energy-isolation process;
• planned inspection and preventive maintenance;
• operator authorization and competency assessment;
• separation of people and powered vehicles;
• load limits, rack protection, and stability controls;
• emergency-stop design and restart prevention.

Automation does not eliminate risk. It changes the risk profile toward unexpected motion, software failure, sensor occlusion, maintenance access, and human-robot interaction. Safety-rated controls, validation, guarding, and controlled access remain necessary.

5.2 Packaging as a safety control

Packaging protects products, but it can also create hazards. A package may be too heavy, difficult to grasp, unstable when stacked, sharp-edged, dusty, chemically incompatible, or impossible to open safely. Packaging design should therefore consider the entire life cycle: filling, palletization, transport, storage, picking, customer opening, reuse, recycling, and disposal.

Important design variables include mass, center of gravity, compression strength, closure integrity, moisture resistance, cushioning, labeling, and compatibility with handling equipment. Standardization can improve safety by making loads predictable, but excessive standardization may create unnecessary void fill or force workers to handle packages that do not fit the product.

For dangerous goods, classification, packaging selection, marking, labeling, documentation, segregation, emergency information, and transport mode requirements must follow the applicable jurisdiction and international or national code. Examples of regulatory regimes include the UN Recommendations on the Transport of Dangerous Goods, national transport regulations, occupational exposure rules, and fire codes. Requirements differ by country and mode; organizations must use current official sources and competent dangerous-goods personnel.

5.3 Product loss and incident prevention

Product loss often signals a safety or process-control weakness. Collapsed pallets, wet cartons, damaged drums, and contaminated goods can injure people as well as reduce inventory accuracy. Prevention includes package testing under realistic distribution hazards, load-pattern standards, stretch-wrap or strapping specifications, humidity controls, inspection of pallets and racks, and clear quarantine procedures.

6. Hazardous-material exposure and emergency preparedness

Hazard communication must enable workers to identify substances, understand hazards, select controls, and respond to releases. A robust program includes an inventory, current safety data sheets, labels, container controls, training, exposure assessment, ventilation where required, spill equipment, waste procedures, and medical or emergency arrangements.

The control sequence remains elimination, substitution, engineering, administrative controls, and PPE. Respiratory protection, for example, cannot compensate for absent ventilation, incompatible storage, or uncontrolled transfer operations. Chemical compatibility is especially important: segregation should be based on actual chemical properties rather than merely product names.

Emergency plans should address fire, spill, collision, power loss, severe weather, cyber disruption, injury, and contamination. They should define authority, communication, evacuation or shelter decisions, isolation, external agency coordination, product disposition, and restart criteria. Drills should test assumptions and reveal dependencies.

7. Resilience and continuity in material-handling systems

A resilient supply chain is not simply one with surplus inventory. Resilience is designed through visibility, redundancy where justified, adaptability, recovery capability, and learning. The goal is to maintain critical service while avoiding wasteful duplication.

7.1 Business impact analysis

A business impact analysis identifies critical processes, their maximum tolerable disruption, dependencies, resources, and recovery priorities. For a distribution center, critical functions may include receiving, order release, picking, packing, shipping, hazardous-goods control, information systems, utilities, and emergency access.

Two useful planning terms are:

• Recovery time objective (RTO): the targeted time within which a process or service should be restored.
• Recovery point objective (RPO): the maximum acceptable period of data loss, expressed as time.

These are managerial targets, not guarantees. A process with a two-hour RTO may require spare equipment, manual work instructions, alternate power, trained personnel, and prearranged suppliers.

7.2 Resilience strategies

Relevant strategies include:

• dual sourcing or qualified alternatives for critical packaging and spare parts;
• cross-training and documented standard work;
• modular conveyors, bypass routes, and manual fallback capability;
• geographically or technically diverse information systems;
• emergency inventory of high-risk or long-lead components;
• supplier visibility and disruption monitoring;
• preventive and predictive maintenance;
• physical protection against flood, fire, heat, and seismic hazards;
• scenario exercises and after-action learning.

Redundancy can create new risk if duplicated systems are maintained poorly or if complexity reduces visibility. Resilience design should therefore examine interdependencies, common-cause failures, and the possibility that a backup depends on the same power, network, supplier, or skilled technician as the primary system.

7.3 Safety and resilience are coupled

A safety incident can become a continuity event: a battery fire closes a warehouse, a chemical spill contaminates inventory, or a serious injury triggers investigation and labor disruption. Conversely, continuity pressure can increase safety risk when managers bypass maintenance, extend shifts, or use unqualified substitutes. Resilient leadership explicitly protects safety during recovery rather than treating safety as a negotiable constraint.

8. Data, technology, and artificial intelligence

Sensors can monitor equipment temperature, vibration, battery condition, location, congestion, and environmental conditions. Wearable devices may identify posture or proximity hazards. Computer vision can detect pedestrian–vehicle interactions, missing PPE, or blocked exits. Digital twins can test layout and disruption scenarios before physical implementation.

AI can support anomaly detection, demand-based staffing, route optimization, and predictive maintenance. However, its use requires governance. False positives can create alert fatigue; false negatives can create dangerous confidence. Cameras and wearables raise privacy concerns, especially if data are used for individual discipline rather than system improvement. Training data may reflect historical bias, and automated recommendations may be difficult to explain.

Good practice includes human oversight, validation against field observations, cybersecurity, access control, data minimization, retention limits, transparent purpose, and a process for workers to challenge erroneous records. Technology should augment professional judgment, not transfer accountability to an opaque model.

9. Illustrative case: redesigning a regional fulfillment center

Background

The fictional NorthRiver Fulfillment Center processes consumer appliances for retail customers. Workers manually lift cartons weighing 8–18 kg from pallets to packing benches. Forklifts and pedestrians share two central aisles. The site uses stretch-wrapped mixed pallets, and a battery-charging area is adjacent to packaging-material storage.

Problem

During a six-month review, the center recorded increasing shoulder and lower-back complaints, three near misses involving forklifts, carton damage of 2.8% of outbound units, and a two-hour shutdown after a damaged battery charger overheated. Management initially proposed refresher training and stricter PPE enforcement.

Analysis

A cross-functional team mapped tasks, observed peak and off-peak work, interviewed employees, and reviewed maintenance and damage data. The lifting assessment showed that the heaviest and most frequent cartons were stored below knee height and beyond a comfortable reach. The screening lifting index for a representative task was above 1. The traffic review found that visual barriers and peak-period congestion made administrative right-of-way rules unreliable. Packaging review found mixed pallet layers with incompatible carton footprints, causing lean and compression damage. The battery area lacked adequate separation from combustible packaging stock.

Approach and solution

The team applied the hierarchy of controls:

1. Reduced carton mass for two product families by splitting shipments where customer requirements allowed.
2. Introduced lift tables and pallet-positioning devices to keep cartons near working height.
3. Reconfigured fast-moving stock into golden-zone locations and reduced reach distance.
4. Installed physical pedestrian barriers, controlled crossings, speed governance, and improved sightlines.
5. Standardized pallet patterns and added a load-stability inspection before release.
6. Relocated combustible packaging and improved charging-area controls in accordance with applicable fire and electrical requirements.
7. Added preventive maintenance triggers for chargers, conveyors, and lift devices.
8. Created a disruption playbook with manual picking capacity, alternate charging arrangements, emergency contacts, and restart authorization.
9. Used worker observations and near-miss reporting in weekly improvement meetings, without incentive structures that discouraged reporting.

Outcome

After three months, the site reported fewer manual-handling complaints, no forklift–pedestrian near misses in the redesigned zone, and a reduction in outbound carton damage. The figures are illustrative rather than independently audited, but the case demonstrates an important principle: training was retained, yet the major gains came from changing load geometry, layout, equipment, packaging, and management controls.

Learning points

• Safety problems often originate in process and packaging design.
• Near misses reveal weak barriers before severe harm occurs.
• Ergonomic, quality, maintenance, and continuity data can identify the same underlying causes.
• A resilient solution includes both prevention and recovery capability.
• Worker participation improves hazard recognition and practical acceptance of controls.

10. Governance, ethics, and implementation

Senior managers are responsible for setting risk appetite, allocating resources, and ensuring that performance incentives do not reward unsafe shortcuts. Supervisors translate policy into daily decisions, while workers contribute task knowledge and should have a meaningful right to report hazards and stop work in imminent-danger situations, subject to local law and procedure.

Ethical safety management requires more than avoiding regulatory penalties. It requires protecting contractors, temporary workers, visitors, and communities; communicating hazards honestly; avoiding retaliation; and investigating incidents for learning rather than blame. Supply-chain contracts should define safety expectations, incident reporting, audit rights, emergency coordination, and responsibilities for packaging and dangerous goods.

Implementation should proceed through a controlled improvement cycle:

1. Establish baseline injury, near-miss, damage, downtime, maintenance, and exposure data.
2. Engage operators and technical specialists in hazard analysis.
3. Prioritize high-consequence and high-exposure risks.
4. Pilot engineering and packaging changes.
5. Validate effectiveness under realistic peak conditions.
6. Update procedures, training, drawings, maintenance plans, and procurement specifications.
7. Monitor leading indicators such as barrier inspections, overdue corrective actions, ergonomic exposures, and near-miss quality.
8. Review lagging indicators such as injuries, releases, damage, downtime, and recovery performance.
9. Reassess after changes, incidents, new products, new technology, or abnormal events.

Challenges and limitations

Risk assessment contains uncertainty. Rare events provide little statistical data; severity judgments can be inconsistent; risk matrices may create false precision; and workers may underreport incidents when trust is low. Ergonomic tools are screening instruments, not substitutes for professional evaluation. Resilience investments can be difficult to justify because their benefit is the disruption that does not occur. Finally, standards and laws vary by jurisdiction and change over time, so current official requirements must be verified before implementation.

Chapter Summary

Safe and resilient material-handling systems are designed rather than achieved through reminders alone. Ergonomic analysis considers force, posture, repetition, duration, vibration, and environmental conditions. The NIOSH Revised Lifting Equation can screen selected two-handed lifting tasks, but its assumptions and limitations must be respected. Risk analysis should identify hazards across normal and abnormal conditions, use transparent prioritization, and examine preventive and mitigative barriers. The hierarchy of controls favors elimination, substitution, and engineering solutions over administrative controls and PPE alone.

Packaging affects worker safety, product protection, hazardous-material control, and continuity. Equipment safety depends on guarding, energy isolation, maintenance, traffic separation, competent operation, and validated automation. Resilience requires business-impact analysis, recovery objectives, alternate capability, cross-training, visibility, and protection against common-cause failures. Technology and AI can improve detection and planning, but they require privacy, cybersecurity, validation, human oversight, and ethical governance. The strongest systems integrate worker participation, operational learning, safety leadership, and continuous verification.

Key Terms

Hazard: A source or situation with potential to cause harm.

Risk: The combination of the likelihood and severity of harm.

Ergonomics: The discipline of designing systems to fit human capabilities and limitations.

Musculoskeletal disorder: An injury or disorder affecting muscles, tendons, nerves, joints, or related structures.

Lifting index: Load weight divided by the recommended weight limit in the NIOSH lifting assessment.

Recommended weight limit: A screening value calculated for specified lifting conditions.

Hierarchy of controls: A ranking of hazard controls from elimination to PPE.

Engineering control: A physical design feature that isolates people from hazards.

Administrative control: A rule, procedure, schedule, training activity, or supervision practice.

Near miss: An event with potential for harm that caused no actual injury or damage.

Bow-tie analysis: A barrier-based method linking causes, a hazardous event, and consequences.

Risk-priority number: The product of severity, occurrence, and detectability ratings in an FMEA context.

Dangerous goods: Materials regulated because of hazards during handling or transport.

Business impact analysis: A method for identifying critical processes, dependencies, and disruption consequences.

Recovery time objective: The targeted time for restoring a process or service.

Recovery point objective: The maximum acceptable period of data loss.

Resilience: The capability to withstand, adapt to, and recover from disruption.

Common-cause failure: A failure that disables supposedly separate controls or systems through a shared cause.

Leading indicator: A measure of preventive activity or control condition.

Lagging indicator: A measure of harm or loss that has already occurred.

Review Questions

Short Answer Questions

1. Distinguish a hazard from risk and provide one material-handling example of each.
2. Name the five levels in the hierarchy of controls.
3. What are the principal limitations of the NIOSH Revised Lifting Equation?
4. Define recovery time objective and recovery point objective.
5. Why can PPE not normally be the primary control for a serious hazard?

Long Answer Questions

1. Explain how packaging design can simultaneously affect ergonomics, product damage, hazardous-material control, and resilience.
2. Describe a complete risk-assessment process for a warehouse that uses forklifts, conveyors, and manual picking.
3. Critically evaluate risk matrices and risk-priority numbers as management tools.
4. Explain how a distribution center can design resilience without creating excessive and costly redundancy.
5. Discuss the ethical and governance issues associated with AI-enabled safety monitoring.

Analytical / Application Questions

1. A worker lifts a 12 kg carton under the conditions in the worked example. Interpret the lifting index and recommend three controls, ranking them by the hierarchy of controls.
2. Design a bow-tie analysis for a pallet collapse during forklift transport, identifying at least three preventive and three mitigative barriers.
3. A site’s only automated sorter fails during a peak period. Develop a continuity response that protects safety while meeting a two-hour recovery target. Identify dependencies and possible common-cause failures.

Multiple Choice Questions with Answers

Question 1

Which intervention is highest in the hierarchy of controls?

A. Providing lifting gloves

B. Delivering refresher training

C. Eliminating the manual lift

D. Posting a warning sign

Correct Answer: C. Eliminating the manual lift

Question 2

In the NIOSH lifting example, a 12 kg load and an RWL of approximately 8.17 kg produce which lifting index?

A. 0.68

B. 1.47

C. 8.17

D. 20.17

Correct Answer: B. 1.47

Question 3

What does a lifting index above 1 generally indicate in the chapter’s screening approach?

A. The task requires no further review

B. The load is automatically illegal

C. Engineering redesign should be considered

D. PPE alone is sufficient

Correct Answer: C. Engineering redesign should be considered

Question 4

Which is a preventive barrier in a forklift–pedestrian bow-tie analysis?

A. Post-incident medical treatment

B. Physical segregation of traffic

C. Product quarantine after impact

D. Business recovery after closure

Correct Answer: B. Physical segregation of traffic

Question 5

What does an RTO specify?

A. Maximum acceptable data loss

B. Target time to restore a process or service

C. Required packaging compression strength

D. Frequency of ergonomic observation

Correct Answer: B. Target time to restore a process or service

Question 6

Which statement about risk-priority numbers is most accurate?

A. They are universal measures of probability

B. They eliminate the need for professional judgment

C. They are useful for comparison but can conceal high-severity risks

D. They apply only to manual lifting

Correct Answer: C. They are useful for comparison but can conceal high-severity risks

Question 7

Which packaging feature most directly reduces unstable pallet loads?

A. A standardized and validated pallet pattern

B. A larger warning label alone

C. Higher picking speed

D. Longer employee shifts

Correct Answer: A. A standardized and validated pallet pattern

Question 8

Why is administrative control generally less reliable than engineering control?

A. It cannot be documented

B. It depends more heavily on consistent human behavior

C. It always costs more

D. It applies only to chemicals

Correct Answer: B. It depends more heavily on consistent human behavior

Question 9

Which is an example of a common-cause failure?

A. Two independent suppliers use different routes

B. A backup conveyor has separate power

C. Primary and backup systems both lose operation because they share one network switch

D. A worker reports a near miss

Correct Answer: C. Primary and backup systems both lose operation because they share one network switch

Question 10

Which governance practice is appropriate for AI-based safety monitoring?

A. Use model outputs without human review

B. Retain all worker data indefinitely

C. Validate performance, protect privacy, and provide human oversight

D. Use alerts primarily for individual punishment

Correct Answer: C. Validate performance, protect privacy, and provide human oversight