Governance, Ethics, Sustainability and Emerging Directions

Introduction

The Internet of Things (IoT) has developed from a collection of connected sensors into a socio-technical infrastructure that influences transport, healthcare, energy, manufacturing, agriculture, homes and public administration. Its value arises from continuous data collection, distributed computation, automated decision-making and interaction between physical and digital systems. These same characteristics create risks that cannot be solved by technical design alone. An insecure medical sensor, a biased predictive model, an opaque smart-city platform or an unsustainable network of disposable devices can produce serious social, economic and environmental harm.

IoT governance therefore concerns the rules, responsibilities, institutions and technical controls through which connected systems are designed, deployed and operated. Ethics asks whether those systems respect human dignity, autonomy, fairness, safety and the public interest. Sustainability examines their effects across the life cycle, including extraction of materials, manufacturing, energy use, maintenance and disposal. Emerging directions such as ambient intelligence, autonomous IoT, 6G, satellite connectivity, blockchain and quantum-safe security expand the opportunity—and the governance challenge.

This chapter treats IoT as a system extending from sensors and actuators to cloud platforms, organizations, regulators, users and affected communities. The objective is not merely to identify risks, but to develop practical methods for responsible innovation.

Learning Objectives

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

1. Explain the principal governance, regulatory and accountability issues in IoT.

2. Evaluate data ownership, consent, privacy, surveillance and algorithmic bias.

3. Apply safety, accessibility, security and standards principles to IoT projects.

4. Assess the environmental impact and circularity of connected products.

5. Describe ambient intelligence, autonomous IoT, 6G, satellite IoT, blockchain and quantum-safe security.

6. Critically identify research challenges and propose responsible IoT strategies.

1. IoT as a Governed Socio-Technical System

A conventional information system usually has identifiable users and an explicit interaction point. IoT systems are more diffuse. A smart building may collect occupancy, temperature, movement and access data through devices owned by different parties. Data can pass through a gateway, a telecommunications network, an analytics platform and several service providers before influencing a physical action. Responsibility is consequently distributed.

A useful governance view considers five layers:

• Device layer: sensors, actuators, embedded software, hardware identity and physical safety.
• Connectivity layer: local networks, cellular systems, satellite links, routing and communications security.
• Data and platform layer: storage, processing, interoperability, APIs, artificial intelligence and access control.
• Organizational layer: procurement, operating procedures, contracts, risk management and incident response.
• Societal layer: law, public policy, cultural expectations, equality, environmental effects and democratic accountability.

A failure at one layer can propagate through the others. For example, inaccurate air-quality sensors can create faulty analytics, which may cause inappropriate public-health decisions. Governance must therefore include technical assurance, organizational controls and mechanisms for public accountability.

Important governance principles include accountability, transparency, proportionality, security by design, privacy by design, inclusiveness, sustainability and contestability. Contestability means that an affected person or organization can question, appeal or seek review of an automated decision.

2. Regulation, Data Rights and Accountability

2.1 The regulatory landscape

IoT is governed by overlapping legal domains rather than by one universal law. These may include data protection and privacy law, cybersecurity obligations, telecommunications regulation, product safety, consumer protection, medical-device regulation, sector-specific critical-infrastructure rules, environmental law, intellectual property and contract law. Requirements vary across jurisdictions, making cross-border deployment complex.

A mature compliance process begins with a legal and risk inventory. The organization should identify:

1. What data are collected and whether they can identify a person or household.

2. The purposes for which data are collected and whether those purposes are compatible with later uses.

3. Which parties determine the purposes and means of processing.

4. Where data are stored, transferred and backed up.

5. Which safety, cybersecurity and sectoral regulations apply.

6. How long data are retained and how they are securely deleted.

7. How individuals can access, correct, restrict or challenge processing where applicable.

Compliance is not equivalent to ethical legitimacy. A practice may satisfy a formal notice requirement while remaining unfair, excessive or incomprehensible to users.

2.2 Data ownership and stewardship

The phrase “data ownership” is often used imprecisely. Data may be generated by an individual, collected by a device manufacturer, processed by a platform and used by a purchaser. Legal rights may include ownership of a physical device, database rights, intellectual property, contractual access rights and data-protection rights. These are not interchangeable.

A better operational approach is data stewardship. A stewardship agreement specifies:

• the categories and provenance of data;
• permitted purposes and prohibited secondary uses;
• access rights for users, operators and researchers;
• retention and deletion rules;
• responsibilities for accuracy and security;
• rules for sharing, aggregation and commercialization;
• procedures for breach notification and dispute resolution.

Data minimization is especially important in IoT. A device should collect data that are adequate and relevant to a defined purpose, rather than recording everything because storage is inexpensive. Edge processing can reduce the transfer of raw personal data by converting it into local, less identifying results.

2.3 Accountability across the value chain

Accountability requires more than naming a privacy officer. It requires evidence that risks were identified, controls were implemented and decisions can be reconstructed. Useful mechanisms include system registers, data-flow maps, threat models, algorithmic impact assessments, supplier audits, software bills of materials, access logs and incident reports.

Contracts should allocate responsibility for patching, vulnerability disclosure, service availability, data breaches, model errors, end-of-life support and recovery after an incident. Liability clauses cannot eliminate the ethical duty of an organization to prevent foreseeable harm.

A practical accountability matrix assigns each important control an owner, an approver, a consulted party and an informed party. This prevents the common situation in which the manufacturer assumes the operator will patch the product while the operator assumes that the manufacturer will do so.

3. Privacy, Surveillance and Human Autonomy

IoT changes privacy because observation becomes continuous, embedded and often invisible. A microphone, camera, wearable, vehicle or smart meter can reveal routines even when the user is not actively interacting with it. Inferences may be more sensitive than the original observations: energy data can suggest occupancy; location traces can reveal relationships; physiological data can indicate health conditions.

Privacy should be analyzed through several dimensions:

• Informational privacy: control over collection, use and disclosure.
• Physical and spatial privacy: freedom from intrusive observation.
• Decisional privacy: freedom to make personal choices without inappropriate manipulation.
• Associational privacy: protection of relationships, movements and group membership.

Consent is valuable but limited. It may be invalid where users have no realistic alternative, cannot understand complex processing, or are not the person being observed. Smart-home systems, for example, can collect information about visitors and children who never agreed to the service. Ethical design therefore combines consent with purpose limitation, privacy-preserving defaults, local processing, strong access control and meaningful opt-out mechanisms.

Surveillance risk is heightened in workplaces, schools and public spaces because power is unequal. Continuous monitoring may create chilling effects, reduce autonomy and normalize social control. A proportionality assessment should ask whether the objective is legitimate, whether the same objective can be achieved with less intrusive data, whether the measure is effective, and whether safeguards and independent oversight exist.

Technical privacy measures include encryption in transit and at rest, pseudonymization, aggregation, differential privacy, secure multi-party computation and federated learning. None is automatically sufficient. Pseudonymized data can sometimes be re-identified, and federated learning can still leak information through model updates. Controls must be evaluated against realistic threat models.

4. Algorithmic Bias, Transparency and Explainability

IoT increasingly uses machine learning for predictive maintenance, traffic management, access control, health monitoring and energy optimization. Bias can enter through non-representative training data, faulty sensors, unequal deployment, inappropriate labels, proxy variables or feedback loops.

For example, a fall-detection system trained mainly on young adults may perform poorly for older people or users with mobility aids. A smart policing system trained on historical enforcement data may reproduce earlier patterns of unequal surveillance. Bias is therefore a system property, not merely a defect in the algorithmic code.

An effective assessment examines:

1. Representation: Are relevant populations and environments represented?

2. Measurement: Do sensors measure the target concept accurately for all groups?

3. Performance: Are error rates, calibration and failure modes acceptable across groups?

4. Impact: Who receives benefits, and who bears false alarms, denial or surveillance?

5. Governance: Can decisions be explained, challenged and corrected?

Fairness criteria can conflict. Equal false-positive rates, equal false-negative rates and equal predictive accuracy cannot generally all be achieved when base rates differ. Consequently, fairness requires a context-sensitive normative decision, not blind selection of a statistical metric.

Explainability should match the decision and its consequences. A safety engineer may need a trace of sensor values, model version and threshold; a patient may need a comprehensible explanation and a human review route. Documentation should record training data characteristics, intended use, limitations, performance by subgroup, monitoring procedures and change history.

5. Safety, Security and Resilience

IoT security and physical safety are closely connected. A compromised thermostat may be inconvenient, whereas a compromised infusion pump, industrial controller or autonomous vehicle can cause injury. Safety engineering therefore considers hazards, likelihood, severity, exposure and controllability, while cybersecurity considers threats, vulnerabilities, attackers and system resilience.

Responsible design includes:

• unique device identity and secure provisioning;
• hardware-backed key protection where justified;
• authenticated communications and least-privilege authorization;
• secure boot and signed firmware;
• vulnerability disclosure and timely patching;
• network segmentation and anomaly detection;
• fail-safe or fail-operational behavior appropriate to the application;
• manual override and emergency shutdown;
• logging, time synchronization and forensic readiness;
• tested backup, recovery and incident-response procedures.

Safety and security objectives can conflict. A system that immediately shuts down on every communication failure may create danger in a process that requires controlled operation. Conversely, continued autonomous operation during a cyberattack may increase harm. The correct response depends on hazard analysis and the application’s safety integrity requirements.

Security should be maintained throughout the product life cycle. Procurement specifications should require a support period, vulnerability reporting channel, update mechanism, cryptographic agility and a clear end-of-life process. A low purchase price is not evidence of low total cost when insecure devices create operational and liability risks.

6. Accessibility, Inclusion and Standards

IoT systems can improve independence through assistive homes, remote health monitoring and accessible transport. They can also exclude people through small interfaces, voice-only interaction, expensive subscriptions, unreliable connectivity or assumptions about language, literacy, physical ability and cognitive capacity.

Inclusive design should address:

• multimodal interaction, including visual, auditory and tactile feedback;
• adjustable text, contrast, timing and alert intensity;
• compatibility with assistive technologies;
• offline or degraded-mode operation;
• multilingual and plain-language communication;
• affordability and equitable access;
• user testing with diverse populations;
• protection against discriminatory automated decisions.

Standards promote interoperability, safety and market confidence. Relevant standardization activity includes IoT reference architectures, networking and identification, information-security management, privacy engineering, functional safety, accessibility, artificial-intelligence governance and environmental management. Standards are not automatically laws, and compliance with one standard does not guarantee ethical or secure behavior. Organizations should distinguish mandatory regulatory requirements, contractual requirements, voluntary technical standards and internal policies.

Interoperability has governance implications. Open interfaces reduce vendor lock-in and support repair and migration, but they must be secured and versioned. A good interface specification defines semantics, authentication, authorization, error behavior, data formats and deprecation policy—not merely connectivity.

7. Environmental Sustainability and Circular IoT

The environmental impact of IoT includes raw-material extraction, semiconductor fabrication, manufacturing, transportation, electricity consumption, cloud processing, battery replacement and electronic waste. The network effect is significant: millions of low-cost devices can produce substantial cumulative impacts even when each device consumes little energy.

Life-cycle assessment evaluates impacts from raw materials through production, use and end of life. A simplified operational energy estimate is:

E_total = N × (E_manufacture + E_use + E_maintenance + E_end-of-life)

where N is the number of devices and each energy term refers to the chosen life-cycle boundary. This expression is an accounting framework, not a universal physical law; the terms must be measured or estimated consistently, and energy should not be confused with total environmental impact. Carbon emissions may be estimated separately using appropriate regional emission factors.

Sustainable design strategies include:

• low-power communication and event-driven sensing;
• energy harvesting where technically appropriate;
• durable hardware and replaceable batteries;
• modularity, repairability and software support;
• recyclable or less hazardous materials;
• local processing to reduce unnecessary data transfer;
• device refurbishment and secure redeployment;
• take-back systems and responsible recycling;
• procurement based on total life-cycle cost and impact.

Trade-offs must be examined. A more powerful edge processor may increase manufacturing impact but reduce communications energy and extend useful functionality. Battery-free operation may require more infrastructure or limit reliability. Sustainability decisions should therefore use measured life-cycle evidence rather than marketing claims.

8. Emerging Directions

8.1 Ambient intelligence

Ambient intelligence describes environments that sense context and respond proactively with minimal explicit interaction. Examples include homes that adjust lighting and heating, hospitals that detect patient deterioration, and factories that adapt workflows. Its promise is seamless assistance; its risk is invisible profiling and loss of user awareness.

Ambient systems should expose their sensing status, provide understandable controls and preserve user agency. Context inference must be conservative where errors have serious consequences. The principle of “calm technology” is useful: systems should demand attention only when necessary and should not make users continuously monitor the technology.

8.2 Autonomous IoT

Autonomous IoT combines sensing, machine learning, communications and actuation to make decisions with limited human intervention. Autonomy can reduce latency and improve operation in hazardous or remote settings, but it creates questions about delegation, verification and liability.

A robust autonomous architecture defines an operational design domain, permitted actions, confidence thresholds, escalation rules and safe fallback states. Human oversight must be meaningful rather than nominal: a person must have sufficient information, time and authority to intervene. Continuous monitoring should detect distribution shift, sensor failure and unexpected behavior.

8.3 6G and integrated sensing and communications

Future 6G systems are expected to support highly integrated terrestrial, aerial and non-terrestrial networks, extremely low latency in selected applications, advanced positioning, network intelligence and joint communication-and-sensing. These are research directions rather than uniformly available capabilities. Their governance issues include energy demand, spectrum allocation, infrastructure concentration, data sovereignty and the privacy implications of networks that infer location and physical context.

Joint sensing and communications may allow a network to estimate movement or occupancy without a conventional camera. This can support safety and automation, but it also expands the amount of environmental information that networks can infer. Privacy safeguards must therefore address inferred data, not only explicit payloads.

8.4 Satellite IoT

Satellite IoT extends connectivity to oceans, deserts, remote infrastructure and disaster zones. It supports environmental monitoring, logistics, agriculture and asset tracking. Constraints include link delay, limited power, radiation exposure, spectrum coordination, terminal cost and difficult physical maintenance. Store-and-forward operation, compact messages and strong device authentication are often important.

Satellite connectivity should complement, not automatically replace, terrestrial networks. Redundancy can improve resilience, but dependence on a single provider creates concentration risk. Operators should plan for service interruption, orbital or ground-segment failures and secure key management across intermittent links.

8.5 Blockchain and distributed ledgers

Distributed ledgers can provide tamper-evident records, shared state and automated transactions among organizations that do not fully trust one another. Potential IoT applications include provenance, maintenance records, decentralized identity and machine-to-machine settlement.

A ledger does not prove that a sensor measurement is true; it records what was submitted. The “oracle problem” remains: trusted mechanisms are needed to connect physical events to digital records. Additional concerns include throughput, latency, governance of permissioned networks, privacy, key loss, energy use and the permanence of incorrectly recorded personal data. Blockchain is appropriate only where shared, append-only coordination has a clear advantage over a conventional database.

8.6 Quantum-safe security

Large-scale quantum computers could threaten widely used public-key systems based on factoring or discrete logarithms. IoT devices often have long lifetimes, limited memory and difficult update paths, creating a “harvest now, decrypt later” concern for data requiring long-term confidentiality.

Quantum-safe migration involves inventorying cryptographic dependencies, identifying data with long confidentiality periods, selecting standardized post-quantum algorithms, testing performance on constrained devices and supporting cryptographic agility. Hybrid mechanisms may be used during transition, but designs must be carefully reviewed for implementation and protocol weaknesses. Post-quantum cryptography does not solve poor authentication, insecure firmware or compromised endpoints; it addresses a particular class of cryptanalytic threat.

9. Research Challenges

Important research directions include:

1. Composable assurance: methods that jointly analyze safety, security, privacy and machine-learning uncertainty.

2. Federated and edge intelligence: efficient learning with verifiable privacy and robustness against poisoned updates.

3. Digital twins: trustworthy synchronization between physical assets and models, including uncertainty representation.

4. Semantic interoperability: shared meaning across vendors, domains and languages.

5. Long-lived security: updateable hardware and cryptographic agility for devices deployed for decades.

6. Sustainable intelligence: measuring the full environmental cost of sensing, communication, computation and replacement.

7. Human autonomy: interfaces and governance for systems that act proactively without becoming coercive.

8. Resilient non-terrestrial networks: secure integration of terrestrial, aerial and satellite systems.

9. Auditable autonomy: evidence-based validation of models under changing environments.

10. Equitable innovation: methods for involving affected communities in requirements, evaluation and oversight.

Research should combine engineering experiments with legal analysis, social research, life-cycle assessment and participatory methods. Technical performance alone is an insufficient measure of success.

10. Applied Case: Connected Cold-Chain Monitoring

Background

The following is an illustrative case. A pharmaceutical distributor proposes battery-powered IoT sensors to monitor temperature, location and door openings in vaccine containers during transport. Data are transmitted through cellular and satellite links to a cloud platform. Machine learning predicts delivery delays and identifies containers at risk.

Problem

The distributor must protect vaccine quality while addressing privacy, cybersecurity, reliability, cost and environmental impact. Drivers may object to continuous location monitoring. Rural routes may have intermittent connectivity. A false temperature alarm could cause unnecessary disposal, whereas a missed alarm could compromise public health.

Analysis

The safety objective is preservation of the required temperature range, while security objectives include preventing false records and unauthorized control. The privacy assessment distinguishes container location from driver-identifying information and limits retention of precise route data. The data-flow map identifies the sensor manufacturer, connectivity provider, distributor, logistics contractors, cloud operator and health authority.

The machine-learning model is evaluated separately for climates, vehicle types, route conditions and sensor models. Its output is advisory: a trained operator reviews high-risk cases. Model version, sensor calibration, battery state, signal quality and relevant readings are retained to support investigation.

Approach and solution

The design uses device identity, signed firmware, encrypted communications, local buffering during outages and a fail-safe alarm when temperature exceeds a defined threshold. Sensors are calibrated and periodically checked against reference instruments. Access is role-based, and route data are pseudonymized for analytics. Contracts specify patch support, incident notification, data deletion and end-of-life take-back.

A dashboard indicates confidence and missing data rather than presenting every prediction as fact. Satellite communication is used only when terrestrial connectivity is unavailable, reducing cost and energy use. Batteries are replaceable, and the sensor housing is designed for refurbishment.

Outcome

The proposed system improves traceability and enables earlier intervention, but it does not eliminate human responsibility. The distributor retains procedures for manual inspection, quarantine and release decisions. Performance and environmental indicators are reviewed periodically, including false alarms, missed events, battery replacement, device recovery and equitable service across remote routes.

Learning points

The case demonstrates that responsible IoT is a system-design activity. Privacy, safety, security, sustainability and accountability must be specified together. Automation should support expert judgment in high-consequence contexts, and every data stream should have a defined purpose, owner, retention period and response procedure.

Best-Practice Checklist

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

• Is the purpose necessary, specific and communicated clearly?
• Are data collection, retention and sharing minimized?
• Are affected people identified, including non-users and vulnerable groups?
• Are safety hazards and cyber threats assessed together?
• Are identity, updates, logging and incident response operational?
• Can users understand, control and challenge important decisions?
• Has performance been tested across relevant populations and environments?
• Are accessibility and offline operation addressed?
• Are interoperability, supplier obligations and end-of-life support documented?
• Has life-cycle environmental impact been estimated and reduced?
• Is there a named accountable owner with authority and resources?

Chapter Summary

IoT governance extends beyond device security to include law, organizational responsibility, human rights, safety, inclusion and environmental stewardship. Data ownership is best operationalized through clear stewardship rules governing purpose, access, retention and accountability. Continuous sensing creates surveillance risks, while machine learning can reproduce bias through data, measurement and deployment practices. Safety-critical IoT requires secure life-cycle engineering, fail-safe behavior, human oversight and tested incident response.

Accessibility, interoperability and standards determine whether IoT benefits are broadly available and sustainable. Life-cycle thinking is essential because manufacturing, energy consumption, maintenance and electronic waste may dominate environmental effects. Ambient intelligence and autonomous IoT offer proactive assistance but require transparency, contestability and limits on delegated authority. 6G, satellite IoT, blockchain and quantum-safe security create new capabilities while introducing new governance and research problems. Responsible IoT is therefore an interdisciplinary practice based on evidence, proportionality, participation and continuous assurance.

Key Terms

Governance: Rules, responsibilities and institutions that direct and control IoT systems.

Data stewardship: Responsible management of data across its collection, use, sharing and disposal.

Accountability: The duty to justify decisions and accept responsibility for outcomes.

Data minimization: Limiting data collection to what is necessary for a defined purpose.

Privacy by design: Integrating privacy protections into system architecture and life-cycle processes.

Surveillance: Systematic observation or inference about people, places or activities.

Algorithmic bias: Systematic and unjustified differences in automated outcomes or errors.

Contestability: The ability to question, appeal or obtain review of a decision.

Explainability: The provision of understandable reasons or evidence for an automated result.

Operational design domain: The conditions within which an autonomous system is intended to operate.

Fail-safe: A behavior that moves toward a safe state when specified failures occur.

Interoperability: The ability of systems to exchange and meaningfully use information.

Life-cycle assessment: Evaluation of environmental impacts across a product or service life cycle.

Circular IoT: IoT designed for durability, repair, reuse, refurbishment and recycling.

Ambient intelligence: Context-aware computing embedded in environments and acting proactively.

Satellite IoT: IoT connectivity provided partly through satellite communication systems.

Distributed ledger: A replicated record maintained among multiple participating entities.

Cryptographic agility: The ability to replace cryptographic algorithms without redesigning the whole system.

Post-quantum cryptography: Cryptographic methods designed to resist attacks from quantum computers.

Review Questions

Short Answer Questions

1. Why is IoT governance considered socio-technical rather than purely technical?

2. Distinguish data ownership from data stewardship.

3. Identify two ways in which IoT can create surveillance risks.

4. What is meant by cryptographic agility?

5. State two environmental benefits of circular IoT design.

Long Answer Questions

1. Explain how accountability should be distributed among an IoT manufacturer, platform provider and system operator.

2. Discuss the relationship between algorithmic bias, sensor quality and fairness evaluation.

3. Analyze the privacy and autonomy risks of ambient intelligence in a smart city.

4. Explain why safety and cybersecurity must be assessed together in autonomous IoT.

5. Compare the opportunities and limitations of blockchain and conventional databases for IoT applications.

Analytical / Application Questions

1. Design a governance plan for a wearable health-monitoring system used by older adults. Include data stewardship, accessibility, safety, consent and incident response.

2. An industrial IoT deployment uses 20,000 sensors. Each sensor has an estimated manufacturing energy of 18 kWh, annual operational energy of 6 kWh, and a four-year service life. Using the stated life-cycle accounting framework and excluding maintenance and end-of-life energy, estimate total energy for four years and explain at least three limitations of the estimate.

3. A remote agricultural network combines terrestrial and satellite IoT, edge analytics and autonomous irrigation. Propose an architecture and evaluate its risks involving connectivity loss, privacy, sustainability, algorithmic error and long-term security.

Multiple Choice Questions with Answers

Question 1

Which principle requires an IoT system to collect only data relevant to a defined purpose?

A. Data minimization

B. Cryptographic agility

C. Network neutrality

D. Fault masking

Correct Answer: A. Data minimization

Question 2

Why is consent alone insufficient as an IoT privacy safeguard?

A. Consent prevents encryption

B. Users may lack a realistic alternative or others may be observed without consent

C. Consent makes data impossible to delete

D. Consent applies only to satellite links

Correct Answer: B. Users may lack a realistic alternative or others may be observed without consent

Question 3

Which activity most directly supports accountability for an automated IoT decision?

A. Removing all system logs

B. Changing the model without documentation

C. Recording model version, inputs and decision history

D. Disabling human review

Correct Answer: C. Recording model version, inputs and decision history

Question 4

What is a major limitation of blockchain for sensor data?

A. It guarantees that the physical measurement is true

B. It cannot record any transaction

C. It eliminates the need for identity management

D. It records submitted data but does not by itself solve the oracle problem

Correct Answer: D. It records submitted data but does not by itself solve the oracle problem

Question 5

Which design choice best supports inclusive IoT interaction?

A. Voice-only control

B. A single fixed text size

C. Multimodal interaction and compatibility with assistive technologies

D. Mandatory continuous connectivity

Correct Answer: C. Multimodal interaction and compatibility with assistive technologies

Question 6

A system has 20,000 devices. Manufacturing energy is 18 kWh per device and four-year operational energy is 24 kWh per device. Ignoring other terms, what is the total energy?

A. 42 kWh

B. 840,000 kWh

C. 480,000 kWh

D. 1,200,000 kWh

Correct Answer: B. 840,000 kWh

Question 7

What is the principal purpose of post-quantum cryptography?

A. To reduce sensor power consumption

B. To replace all network protocols with blockchain

C. To resist attacks enabled by future quantum computers

D. To guarantee unbiased machine-learning outputs

Correct Answer: C. To resist attacks enabled by future quantum computers

Question 8

In an autonomous IoT system, meaningful human oversight requires that a human have sufficient information, time and what else?

A. Authority to intervene

B. Access to raw data from every device

C. A blockchain wallet

D. Unlimited computing power

Correct Answer: A. Authority to intervene

Question 9

Which statement best describes life-cycle assessment?

A. It measures only electricity used during device operation

B. It evaluates impacts across stages such as materials, manufacturing, use and disposal

C. It measures only software quality

D. It replaces all safety analysis

Correct Answer: B. It evaluates impacts across stages such as materials, manufacturing, use and disposal

Question 10

Why should a high-consequence machine-learning IoT output often be treated as advisory?

A. Machine learning cannot process sensor data

B. Human review can examine uncertainty and exceptional cases before harmful action

C. Advisory outputs require no validation

D. Autonomous systems never need monitoring

Correct Answer: B. Human review can examine uncertainty and exceptional cases before harmful action