Ethics, Governance, Regulation and Sustainability

Introduction

The Internet of Things (IoT) connects sensors, actuators, appliances, vehicles, industrial machines, buildings, medical devices and digital services into socio-technical systems. These systems collect observations about people and environments, make inferences, trigger actions and increasingly employ artificial intelligence (AI). Their value is therefore inseparable from questions of power, responsibility, legality and environmental impact.

An IoT deployment may improve energy efficiency, healthcare, transport and industrial productivity. The same deployment may also enable continuous surveillance, discriminatory decisions, unsafe automation, insecure infrastructures, excessive resource consumption or the transfer of personal data across jurisdictions. Technical functionality alone is consequently an inadequate measure of success. A professionally designed IoT system must be trustworthy, accountable, lawful, secure, inclusive and environmentally responsible.

This chapter treats ethics, governance, regulation and sustainability as related but distinct dimensions. Ethics concerns what ought to be done; governance concerns how decisions, responsibilities and controls are organized; regulation concerns legally enforceable requirements; and sustainability concerns the long-term social, environmental and economic viability of systems.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Explain the ethical characteristics and risks of IoT systems.
2. Distinguish ethics, governance, compliance, regulation and corporate responsibility.
3. Apply privacy, security, safety, fairness and accountability frameworks to IoT projects.
4. Analyze the regulatory life cycle of data and connected products.
5. Evaluate sustainability across the IoT hardware, communication, cloud and disposal life cycle.
6. Design governance mechanisms, impact assessments and controls for an IoT deployment.
7. Critically assess an applied IoT case from technical, legal, social and environmental perspectives.

1. The Socio-Technical Nature of IoT

A conventional information system generally processes data supplied by identifiable users or organizational databases. IoT systems differ in four important ways:

• Ubiquitous observation: sensing may occur continuously, passively and in public or semi-private spaces.
• Distributed agency: decisions are spread across devices, gateways, cloud services, algorithms, operators and third parties.
• Physical consequences: software can open a lock, alter a machine setting, administer treatment or control traffic.
• Extended life cycle: devices may remain deployed for many years, while vendors, standards, ownership and laws change.

These characteristics create an accountability gap. A person affected by an IoT decision may not know which organization collected the data, which algorithm produced the inference, which device acted, or how to challenge the outcome. Responsibility can become fragmented among the manufacturer, software provider, connectivity operator, data controller, system integrator, building owner and user.

IoT governance must therefore cover the complete chain:

person or environment → sensor → device firmware → network → platform → analytics or AI → decision → actuator or service → social and environmental outcome

A governance analysis that examines only the cloud database is incomplete. The sensor may be over-collecting data, firmware may be insecure, a network provider may retain metadata, and an actuator may create physical danger even when the database is protected.

2. Ethical Foundations for IoT

2.1 Autonomy and informed choice

Autonomy requires that people can make meaningful decisions about activities affecting them. IoT challenges autonomy when consent is hidden in lengthy terms, when opting out makes a service unusable, or when data are collected from people who never directly interacted with the system. A workplace location system, for example, may monitor employees who had no realistic ability to refuse.

Meaningful choice requires clear notice, purpose specification, accessible controls and a genuine alternative where feasible. Consent is not always the most appropriate legal or ethical basis: it may be coercive in employment, impossible to obtain from passers-by, or insufficient for safety-critical processing. Organizations must identify a legitimate purpose and select a proportionate basis for processing.

2.2 Privacy and contextual integrity

Privacy is broader than secrecy. It includes control over personal information, freedom from inappropriate observation and the ability to live without unjustified profiling. Helen Nissenbaum’s contextual integrity perspective is especially useful for IoT: information is appropriate when it flows according to the norms of a particular context, including the sender, recipient, subject, type of information and transmission principle.

A smart-home motion sensor illustrates the issue. Residents may accept data use for heating control but reject its transfer to an insurer or landlord. The same data can be ethically acceptable in one context and intrusive in another.

Privacy protection should include:

• data minimization and purpose limitation;
• local processing when cloud transmission is unnecessary;
• aggregation, pseudonymization or anonymization where appropriate;
• retention limits and deletion mechanisms;
• encryption in transit and at rest;
• access controls and auditable data sharing;
• interfaces that reveal sensing and processing clearly.

Pseudonymization reduces direct identifiability but does not eliminate personal-data risk. Location, timing and device identifiers can often be combined with other datasets to re-identify individuals.

2.3 Beneficence, non-maleficence and safety

Beneficence requires producing legitimate benefits, whereas non-maleficence requires avoiding foreseeable harm. In IoT, safety is both a technical and ethical obligation. A connected insulin pump, industrial controller or vehicle must fail safely when communication, power, software or authentication fails.

Safety engineering commonly distinguishes hazards, initiating events, consequences and controls. A hazard analysis should ask:

1. What harmful event could occur?
2. Under what operating conditions?
3. Who or what could be affected?
4. How severe and probable is the harm?
5. Which preventive, detective and mitigative controls exist?
6. What happens when the system is unavailable or compromised?

Security and safety are interdependent. A stolen credential may become a safety hazard when it permits an attacker to alter a robotic process. Conversely, a safety override may create a cybersecurity weakness. A secure design must preserve emergency operation without creating an unrestricted bypass.

2.4 Justice, fairness and inclusion

IoT benefits and burdens may be distributed unequally. Predictive policing sensors may concentrate surveillance in already over-monitored communities. Smart-city services may exclude residents without smartphones or stable connectivity. AI models trained on unrepresentative sensor data may perform poorly for particular populations.

Fairness should be assessed across the system rather than only at the algorithm. Relevant questions include:

• Who is represented in the data?
• Who bears installation, monitoring or false-alarm costs?
• Are accessibility needs addressed?
• Can people contest automated decisions?
• Are benefits available to low-income, rural, elderly and disabled populations?

No single statistical fairness criterion is universally correct. Equal false-positive rates, equal access, equal accuracy and equal outcomes may conflict. The appropriate criterion depends on the decision context, legal requirements and the harms involved.

2.5 Accountability and explainability

Accountability means that an identifiable actor can explain, justify and, where necessary, remedy an action. It is not achieved merely by naming a responsible executive. It requires records, decision rights, monitoring, complaint routes and effective remedies.

Explainability should be proportionate to risk. A simple rule controlling lighting may need an understandable user explanation. A machine-learning system influencing medical triage, insurance or employment requires stronger documentation, validation, human oversight and contestability. Technical opacity does not transfer responsibility to the algorithm.

3. Governance: From Principles to Organizational Control

Governance is the framework through which an organization directs and controls IoT activities. It differs from management: management operates systems and projects, while governance establishes authority, objectives, accountability and oversight.

3.1 Governance layers

An effective model has at least five layers:

Governance should begin before procurement. A project may be technically feasible but unacceptable because its purpose is disproportionate, its data cannot be deleted, or its vendor cannot provide security updates throughout the intended life.

3.2 Roles and responsibility allocation

A responsibility matrix should identify, for every major activity, who is responsible for performing it, accountable for its outcome, consulted before decisions and informed afterward. Typical roles include:

• product owner or system operator;
• data controller and data processor, where applicable;
• device manufacturer and firmware maintainer;
• cloud and connectivity providers;
• information-security and privacy officers;
• safety engineer and domain expert;
• procurement and legal functions;
• affected users and community representatives.

Contracts should specify security baselines, vulnerability disclosure, update duration, audit rights, data location, subcontractors, breach notification, service termination, data return and secure disposal. Vague contractual language such as “industry-standard security” is inadequate unless the standard and evidence are identified.

3.3 Risk and impact assessment

IoT projects should combine several assessments rather than relying on a single checklist:

• Data protection impact assessment: necessity, proportionality, data flows, rights and safeguards.
• Threat and vulnerability assessment: attackers, assets, attack paths and mitigations.
• Safety assessment: hazards, failure modes, safe states and emergency procedures.
• Algorithmic impact assessment: data quality, validity, bias, explainability and human review.
• Environmental life-cycle assessment: materials, energy, emissions, repair and end-of-life effects.
• Human-rights impact assessment: surveillance, discrimination, labor effects and access to remedy.

A useful governance principle is proportionality: the depth of assessment, testing and oversight should increase with the severity, scale, reversibility and likelihood of harm.

3.4 Security governance across the life cycle

Baseline practices include unique device identities, secure boot, signed updates, least privilege, strong authentication, encrypted communication, protected key storage, vulnerability disclosure, asset inventory, logging and incident response. Security must continue after deployment. An inexpensive device with no update mechanism may create a long-lived risk externalized to users and the public internet.

The organization should define an end-of-support policy before procurement. If a device cannot receive security updates, it must be isolated, replaced, retired or subject to compensating controls. “Install and forget” is incompatible with responsible IoT governance.

4. Regulatory and Legal Dimensions

Regulation converts selected social values into enforceable obligations. It may govern privacy, cybersecurity, product safety, telecommunications, consumer protection, accessibility, competition, environmental impacts, employment and sector-specific activities such as healthcare or finance.

4.1 Regulatory categories

Specific duties vary by jurisdiction. Organizations must not assume that a device is outside regulation merely because it is marketed as a consumer gadget. Its function, data, location, users and consequences determine the applicable regime.

4.2 Data life-cycle compliance

Legal analysis should map data from collection through deletion:

1. Collection: Is the data necessary, transparent and lawfully obtained?
2. Transmission: Are communications protected and cross-border transfers controlled?
3. Storage: Are retention periods, access rights and segregation defined?
4. Use: Is secondary use compatible with the original purpose?
5. Disclosure: Are processors, partners and public authorities governed?
6. Inference: Are derived attributes treated as potentially sensitive?
7. Deletion: Can copies, backups and device caches be removed or rendered inaccessible?

Data that appear non-personal in isolation may become personal when combined. A temperature reading tied to a device identifier and time pattern can reveal occupancy, health conditions or work routines.

4.3 Compliance versus responsible practice

Compliance is necessary but insufficient. A practice can be lawful yet ethically unacceptable, especially where legal rules lag behind technology or where individuals have no realistic bargaining power. Conversely, an ethical ambition must be translated into specific requirements and evidence if it is to survive organizational change.

A mature organization maintains a compliance register linked to system requirements, tests and evidence. Evidence may include design records, consent or notice records, penetration-test reports, update logs, supplier attestations, incident records and audit results. Compliance claims should be specific, current and verifiable.

5. Sustainability of IoT Systems

Sustainability includes environmental integrity, social well-being and economic durability. IoT can reduce energy, transport and material waste, but it can also increase consumption through billions of short-lived devices, batteries, sensors, networks and cloud workloads.

5.1 Life-cycle perspective

The relevant system boundary extends from raw-material extraction to manufacturing, transport, deployment, operation, maintenance and end-of-life. Important impacts include energy use, greenhouse-gas emissions, water consumption, critical minerals, hazardous substances, e-waste and labor conditions in supply chains.

Operational efficiency alone can be misleading. A smart thermostat may reduce building energy use while requiring frequent battery replacement and generating rebound effects through increased device ownership. Sustainability claims should therefore use life-cycle evidence rather than a single operational metric.

A basic energy model is:

Here, E_total is life-cycle energy in joules or kilowatt-hours; each term represents the energy attributable to a life-cycle stage. The model is an accounting identity, not a prediction by itself. Boundaries, allocation rules, time period and data sources must be documented. For emissions, a comparable model is:

where A_i is activity amount for source i, such as electricity consumed in kilowatt-hours, and EF_i is the corresponding emission factor, such as kilograms of CO₂-equivalent per kilowatt-hour. Units must be compatible, and emissions from manufacturing, electricity, logistics and disposal should not be omitted without justification.

5.2 Design strategies

Sustainable IoT design includes:

• long service life and modular hardware;
• repairability and replaceable batteries;
• energy-efficient sensing, communication and computation;
• adaptive sampling rather than continuous transmission;
• local analytics where it reduces communication and cloud demand;
• open interfaces that reduce premature obsolescence;
• secure updateability without requiring hardware replacement;
• responsible minerals and supplier due diligence;
• take-back, refurbishment and material recovery.

Trade-offs must be evaluated. Local processing can reduce network energy and privacy exposure, but may increase device computation and replacement complexity. A low-power protocol may have limited bandwidth. Stronger encryption may impose computation and energy costs, though these costs are usually justified by risk. Sustainable design is therefore an optimization problem involving environmental, security, social and performance constraints rather than a search for one universal solution.

5.3 Social and economic sustainability

A system is not sustainable if it displaces workers without transition support, excludes people with disabilities, imposes unaffordable maintenance costs or depends on a vendor likely to abandon the platform. Procurement should evaluate total cost of ownership:

The terms are monetary costs over a stated evaluation period. They should include labor, training, upgrades, security incidents and disposal rather than only the purchase price. Discounting may be used for investment analysis, but the discount rate and time horizon must be reported because they affect comparisons.

6. An Integrated Responsible-IoT Framework

A practical framework can be organized as a continuous cycle:

define purpose → identify stakeholders → map data and physical effects → assess impacts → design controls → test and validate → deploy transparently → monitor → remedy and improve → retire responsibly

Step 1: Define purpose and necessity

State the problem, intended benefit, affected groups and alternatives. Ask whether IoT is necessary or whether a less intrusive solution would achieve the same objective.

Step 2: Map stakeholders and power

Identify users, non-users, workers, bystanders, vulnerable groups, suppliers, regulators and communities. Record who gains, who bears risk and who can challenge decisions.

Step 3: Create data and dependency maps

Document sensors, identifiers, data transformations, algorithms, external services, actuators, jurisdictions and retention points. Include metadata and inferred data, not only raw measurements.

Step 4: Establish requirements

Translate principles into verifiable requirements, such as maximum retention, update support period, fail-safe behavior, accessibility, incident notification time and energy ceiling. A principle without a measurable acceptance criterion is difficult to govern.

Step 5: Conduct multidisciplinary review

Include engineering, cybersecurity, privacy, legal, safety, sustainability, domain and community perspectives. Independent review is particularly important for high-impact systems.

Step 6: Validate before scaling

Use representative data, adversarial tests, usability studies, accessibility checks, safety tests, environmental measurements and pilot deployment. Test failure and misuse conditions, not only normal operation.

Step 7: Operate with transparency and remedy

Provide notices, dashboards, complaint channels, human escalation and correction processes. Monitor drift, new vulnerabilities, changing laws, energy consumption and disparate impacts.

Step 8: Retire and learn

Revoke credentials, erase or return data, disable services, recover equipment and document lessons. Retirement is part of design, not an afterthought.

7. Worked Application: Illustrative Smart-Campus Occupancy System

Background

This illustrative case concerns a university planning a system of room sensors, Wi-Fi analytics and access-control records to optimize heating, ventilation, lighting and classroom allocation. The proposed platform would estimate occupancy and publish room availability.

Problem

The university expects lower energy use and better space utilization. Students and staff, however, are concerned that occupancy data could reveal attendance, religious practice, health appointments or union activity. The system may also misclassify crowded rooms, disadvantage people whose devices are not detected and continue collecting data after the original energy purpose has ended.

Analysis

The stakeholders include students, staff, visitors, facilities personnel, disability services, cybersecurity staff, suppliers and university leadership. The primary risks are:

• privacy: persistent identifiers and movement histories;
• function creep: use for attendance discipline or employee monitoring;
• fairness: inaccurate detection for people without smartphones or with assistive devices;
• security: compromise of access and location data;
• safety: incorrect ventilation or emergency assumptions;
• sustainability: battery replacement and electronic waste;
• governance: unclear ownership between facilities and IT.

A less intrusive alternative is to use anonymous aggregate counts from ceiling sensors, process data locally and transmit only room-level totals at fixed intervals. Exact location histories are unnecessary for the stated energy purpose.

Approach and solution

The university adopts the following controls:

1. Sensors produce aggregate counts and discard raw imagery locally.
2. Device identifiers are not used for room optimization.
3. Data are retained only long enough to calculate operational trends.
4. A separate, documented approval is required for any secondary purpose.
5. Energy and accuracy are tested across room types and accessibility conditions.
6. A public data notice explains sensing, purposes, retention and contacts.
7. Facilities staff receive room-level data; individual movement data are prohibited.
8. Devices use unique credentials, signed updates and network segmentation.
9. Manual overrides preserve ventilation and emergency operation.
10. Procurement requires repair, battery replacement, update support and take-back.
11. An oversight group reviews incidents, complaints, accuracy and energy outcomes each semester.

Outcome

The solution provides operational information without creating a permanent movement database. It does not eliminate risk: aggregate data can still reveal patterns in small rooms, sensors can fail and future administrators may seek new uses. Therefore, governance controls, periodic review and a clear prohibition on repurposing remain essential.

Learning points

The case demonstrates that responsible IoT is not achieved by adding privacy notices after deployment. Purpose limitation influenced the architecture; stakeholder analysis exposed non-obvious harms; local processing reduced both privacy exposure and network dependence; procurement determined long-term security and sustainability; and governance preserved accountability after launch.

8. Contemporary Developments and Critical Issues

Edge AI and federated architectures

Edge processing can reduce latency, bandwidth and exposure of raw data. Federated learning can keep training data distributed, but model updates may still leak information, and participating devices may have unequal computational capacity. These approaches are privacy-enhancing possibilities, not automatic guarantees.

Digital twins and cyber-physical replicas

Digital twins integrate sensor data, simulations and operational decisions. They can improve maintenance but create concentrated models of assets, workers and environments. Governance must distinguish measured facts, estimates, predictions and simulated scenarios, and must prevent users from treating uncertain predictions as reality.

Generative AI and autonomous agents

Generative systems may summarize IoT data, generate control recommendations or interact with operators. Risks include fabricated explanations, prompt injection through sensor data, leakage of sensitive observations and overreliance by human operators. High-impact actions should require constrained tools, authorization, logging, validation and meaningful human oversight.

Standardization and interoperability

Interoperability can reduce vendor lock-in and extend product life, but shared interfaces may enlarge attack surfaces. Standards should address identity, update mechanisms, data semantics, safety behavior and decommissioning, not merely connectivity. Open standards are most beneficial when accompanied by conformance testing and clear responsibility.

The politics of infrastructure

IoT is not neutral infrastructure. Choices about what is sensed, whose data are valued and which outcomes are optimized reflect institutional priorities. A city optimized for traffic flow may reduce pedestrian safety; a factory optimized for throughput may intensify worker surveillance. Ethical review must therefore ask not only “Can the system work?” but also “What social arrangement does it create, for whom and under whose authority?”

9. Challenges and Limitations

First, legal requirements differ across jurisdictions and change over time. Organizations need current legal advice and regulatory monitoring rather than relying on static summaries. Second, privacy-enhancing techniques cannot guarantee anonymity against all auxiliary data. Third, impact assessments can become ceremonial if decision-makers ignore their findings. Fourth, sustainability measurements depend on boundaries, supplier data and assumptions; apparently precise numbers may conceal uncertainty. Fifth, fairness metrics can conflict, and correcting one disparity may create another. Sixth, distributed supply chains make it difficult to verify firmware, labor conditions and material origins. Finally, transparency may conflict with security or intellectual property, so disclosure must be carefully designed rather than treated as absolute.

These limitations reinforce the need for iterative, evidence-based governance. Responsible IoT is not a certification achieved once; it is a continuing organizational capability.

10. Best-Practice Checklist

Before deployment, an IoT team should be able to answer yes to the following questions:

• Is the purpose specific, necessary and proportionate?
• Have affected non-users and vulnerable groups been consulted?
• Are data flows, inferences, third parties and jurisdictions documented?
• Are collection, retention and secondary use limited?
• Are secure defaults, update support and vulnerability response defined?
• Have safety, privacy, fairness, human-rights and environmental impacts been assessed?
• Are automated decisions explainable, contestable and subject to appropriate human oversight?
• Are failure modes, emergency operation and manual overrides tested?
• Are energy, materials, repair and end-of-life requirements included in procurement?
• Are responsibility, audit evidence, incident response and remedies assigned?
• Is there a retirement plan that revokes access and handles data and equipment responsibly?

Chapter Summary

IoT systems are socio-technical infrastructures whose effects extend from data collection into physical, social and environmental domains. Ethical analysis must address autonomy, privacy, safety, beneficence, justice, inclusion, accountability and explainability. Governance converts these concerns into roles, decision rights, impact assessments, technical controls, monitoring and remedies.

Regulation may impose obligations relating to data protection, cybersecurity, product safety, telecommunications, consumer protection, sector-specific risk and environmental responsibility. Compliance is essential but does not exhaust ethical responsibility. Sustainability requires a life-cycle view that includes manufacturing, operation, maintenance and end-of-life, together with social and economic durability.

A responsible-IoT life cycle begins with purpose and stakeholder analysis, maps data and physical dependencies, establishes measurable requirements, validates controls, operates transparently and retires systems safely. The smart-campus case shows how architectural choices, procurement and governance can reduce harm while preserving useful functionality. Emerging edge AI, digital twins and autonomous systems increase both opportunity and the need for oversight.

Key Terms

Ethics: The study and application of principles concerning right conduct and human well-being.

Governance: The structures and processes used to direct, control and hold an organization accountable.

Regulation: Legally enforceable rules governing conduct, products, data or institutions.

Accountability: The duty and ability to explain decisions, accept responsibility and provide remedy.

Autonomy: A person’s capacity to make informed and voluntary choices.

Contextual integrity: The view that privacy depends on appropriate information flows within a social context.

Data minimization: Collecting and retaining only data necessary for a defined purpose.

Pseudonymization: Replacing direct identifiers while recognizing that re-identification may remain possible.

Purpose limitation: Restricting data use to specified and compatible purposes.

Privacy impact assessment: A structured evaluation of privacy risks and safeguards.

Threat modelling: Systematic identification of assets, adversaries, vulnerabilities and attack paths.

Safety case: A reasoned, documented argument supported by evidence that a system is acceptably safe.

Fairness: The requirement that benefits, burdens and decisions do not impose unjustified discrimination.

Explainability: The capacity to provide understandable reasons for system outputs or actions.

Life-cycle assessment: Evaluation of environmental impacts across a product or system’s life stages.

E-waste: Discarded electrical and electronic equipment and its components.

Total cost of ownership: The complete cost of acquiring, operating, maintaining and retiring a system.

Security update: A software or firmware change that addresses vulnerabilities or strengthens protection.

Data controller: An organization that determines purposes and means of processing, where defined by applicable law.

Responsible innovation: Innovation designed and governed to anticipate impacts, include stakeholders and support public benefit.

Review Questions

Short Answer Questions

1. Why is IoT accountability often fragmented across several organizations?
2. Distinguish pseudonymization from anonymization.
3. What is the purpose of a data protection impact assessment?
4. Why must IoT sustainability be evaluated across the life cycle?
5. Explain why compliance does not necessarily equal ethical acceptability.

Long Answer Questions

1. Discuss autonomy, privacy, safety, justice and accountability as ethical foundations for IoT.
2. Explain the layers of IoT governance and show how they interact.
3. Analyze how data protection, product cybersecurity and product safety regulation overlap in a connected medical device.
4. Evaluate the environmental and economic trade-offs of edge processing compared with cloud processing.
5. Design a responsible-IoT governance framework for a smart-city deployment.

Analytical / Application Questions

1. A university proposes wearable badges that continuously record student location to improve campus services. Conduct a stakeholder, privacy, fairness and governance analysis and recommend an alternative architecture.
2. A connected industrial controller receives a security update that may temporarily disable a safety function. Develop a decision process balancing security, safety, continuity and accountability.
3. Compare two IoT procurement proposals: one has a lower purchase price but three years of support; the other costs more but offers ten years of updates, repairable hardware and vendor take-back. Use total cost of ownership, risk and sustainability reasoning to recommend one.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes ethics from regulation in IoT?

A. Ethics concerns only software quality, while regulation concerns only hardware.

B. Ethics concerns what ought to be done, while regulation establishes legally enforceable requirements.

C. Ethics is voluntary law, while regulation is always international.

D. Ethics applies after deployment, while regulation applies before deployment only.

Correct Answer: B. Ethics concerns what ought to be done, while regulation establishes legally enforceable requirements.

Question 2

Which design choice most directly supports data minimization in a smart-building system?

A. Storing raw camera footage indefinitely

B. Sending every sensor reading to multiple vendors

C. Processing occupancy counts locally and transmitting only aggregate totals

D. Linking room readings to permanent personal identifiers

Correct Answer: C. Processing occupancy counts locally and transmitting only aggregate totals

Question 3

Why does pseudonymization not eliminate privacy risk?

A. Pseudonymized data cannot be encrypted

B. Pseudonyms always reveal a person’s name

C. Pseudonymized records may be re-identified using auxiliary information

D. Pseudonymization prevents access control

Correct Answer: C. Pseudonymized records may be re-identified using auxiliary information

Question 4

Which activity is primarily part of safety analysis?

A. Identifying hazards, consequences and safe failure states

B. Choosing a marketing slogan

C. Maximizing data retention

D. Removing all human oversight

Correct Answer: A. Identifying hazards, consequences and safe failure states

Question 5

In the life-cycle energy expression E_total = E_manufacture + E_transport + E_operation + E_maintenance + E_end-of-life, what does E_operation represent?

A. Energy used only to manufacture sensors

B. Energy associated with running the deployed system

C. Energy used to recycle discarded equipment only

D. Energy embedded in raw materials only

Correct Answer: B. Energy associated with running the deployed system

Question 6

Which governance mechanism most directly assigns decision authority and accountability among IoT participants?

A. A responsibility matrix

B. A higher sampling frequency

C. A larger storage cluster

D. A device color standard

Correct Answer: A. A responsibility matrix

Question 7

A connected device has no mechanism for security updates. What is the most significant governance concern?

A. It may become a long-lived vulnerability after deployment

B. It will necessarily use too much display space

C. It cannot collect any data

D. It automatically becomes compliant with all regulations

Correct Answer: A. It may become a long-lived vulnerability after deployment

Question 8

Which statement about sustainability is most accurate?

A. Operational energy is the only relevant environmental impact

B. A low purchase price always indicates sustainability

C. Sustainability requires considering environmental, social and economic effects across the life cycle

D. Sustainability and cybersecurity are unrelated

Correct Answer: C. Sustainability requires considering environmental, social and economic effects across the life cycle

Question 9

Why should IoT systems provide human contestability for high-impact automated decisions?

A. To guarantee that algorithms never make errors

B. To allow affected people to seek explanation, correction or remedy

C. To eliminate the need for system testing

D. To permit unlimited secondary data use

Correct Answer: B. To allow affected people to seek explanation, correction or remedy

Question 10

Which procurement requirement best supports responsible retirement of IoT equipment?

A. A prohibition on documenting device ownership

B. Vendor take-back, data deletion and credential revocation procedures

C. Permanent retention of all device credentials

D. Removal of all update obligations

Correct Answer: B. Vendor take-back, data deletion and credential revocation procedures