IB DP · PastPaper.sampleTitle

MetadataPastPaper.sampleTitle

Thinka Nov 2025 HL IB Diploma Programme-Style Mock — Digital society

82 PastPaper.marks210 PastPaper.minutes2025
PastPaper.analysisPastPaper.samplePaper
An original Thinka practice paper modelled on the structure and difficulty of the Nov 2025 HL IB Diploma Programme Digital society paper. Not affiliated with or reproduced from IB.

Paper 1 Section A

Answer two structured questions out of three. Each question is worth 20 marks and focuses on digital concepts, content, and contexts.
14 PastPaper.question · 40 PastPaper.marks
PastPaper.question 1 · Short Answer
2 PastPaper.marks
State two human rights concerns related to the deployment of mass facial recognition surveillance in public spaces.
PastPaper.showAnswers

PastPaper.workedSolution

Mass facial recognition surveillance presents significant challenges to internationally recognized human rights:
1. **Right to privacy**: Constant surveillance in public spaces without consent systematically tracks individuals' movements, eliminating anonymity.
2. **Freedom of assembly**: Knowing that they are being monitored and identified can create a 'chilling effect,' discouraging citizens from participating in peaceful protests, demonstrations, or public gatherings.

PastPaper.markingScheme

Award [1 mark] for each valid human rights concern identified, up to a maximum of [2 marks].

Suitable responses include:
- Violation of the right to privacy / loss of public anonymity.
- Chilling effect on the freedom of assembly and association.
- Risk of discrimination or profiling (e.g., algorithmic bias against minority groups).
- Threat to the presumption of innocence (being subjected to constant police monitoring without suspicion of a crime).
PastPaper.question 2 · Short Answer
2 PastPaper.marks
Outline how historical bias can become embedded in an artificial intelligence (AI) system used for hiring employees.
PastPaper.showAnswers

PastPaper.workedSolution

Artificial intelligence systems learn from training data. If a company's historical hiring decisions favored a specific demographic group (e.g., due to human bias or systemic inequalities), the training dataset will reflect this skew. The AI algorithm detects these patterns as indicators of 'success' and subsequently ranks applicants matching the historically favored demographic higher, thereby reinforcing and automating the pre-existing bias.

PastPaper.markingScheme

Award [1 mark] for explaining that the AI relies on training datasets consisting of past human hiring decisions which contain existing biases.
Award [1 mark] for explaining how the system codifies/replicates this bias by identifying patterns in the data and using them to evaluate new candidates.
PastPaper.question 3 · Short Answer
2 PastPaper.marks
Identify two physical safety challenges associated with using autonomous drones for package delivery in urban environments.
PastPaper.showAnswers

PastPaper.workedSolution

Deploying autonomous drones in highly populated urban environments poses distinct physical safety risks:
1. **Obstacle avoidance failure**: Urban environments have complex, dynamic obstacles such as power lines, construction cranes, trees, and other drones. A failure in the drone's collision avoidance sensors could lead to crashes.
2. **Malfunction and payload drop**: Technical faults, loss of signal, battery depletion, or strong wind gusts can cause a drone to fall from the sky, risking serious injury to pedestrians or damage to vehicles below.

PastPaper.markingScheme

Award [1 mark] for each distinct physical safety challenge identified, up to a maximum of [2 marks].

Suitable responses include:
- Collisions with physical structures (power lines, buildings, streetlights).
- Injuries to pedestrians from falling drones/payloads due to technical failure.
- Interference or collision with birds or local wildlife.
- Hazard of rotating propeller blades to users retrieving the package.
PastPaper.question 4 · Short Answer
2 PastPaper.marks
Explain how personalized recommendation algorithms can lead to the creation of epistemic bubbles for internet users.
PastPaper.showAnswers

PastPaper.workedSolution

Personalized recommendation algorithms are designed to maximize user engagement by showing content that aligns with the user's past clicks, searches, and likes (selective exposure). Over time, this algorithmic filtering isolates the user from alternative views and diverse information sources. The user is left in an epistemic bubble (or echo chamber) where their existing opinions, biases, and beliefs are continuously reinforced and never challenged.

PastPaper.markingScheme

Award [1 mark] for identifying that algorithms selectively filter and serve content based on past preferences/behavior to maximize engagement.
Award [1 mark] for explaining that this selective filtering systematically excludes opposing viewpoints, resulting in a narrow, self-reinforcing information ecosystem (epistemic bubble).
PastPaper.question 5 · Short Answer
2 PastPaper.marks
State two environmental consequences of planned obsolescence in consumer electronics.
PastPaper.showAnswers

PastPaper.workedSolution

Planned obsolescence—designing products with a artificially limited useful life so they become obsolete quickly—causes significant environmental degradation:
1. **E-waste accumulation**: Rapid replacement cycles lead to millions of tons of obsolete devices being discarded annually, often ending up in landfills where toxic chemicals (like lead and mercury) leach into the ecosystem.
2. **Resource depletion**: Manufacturing new electronics requires continuous extraction of finite raw materials, such as lithium, cobalt, and gold, leading to destructive mining practices and high carbon emissions from production and global shipping.

PastPaper.markingScheme

Award [1 mark] for each valid environmental consequence identified, up to a maximum of [2 marks].

Suitable responses include:
- Increase in toxic electronic waste (e-waste) in landfills.
- Depletion of scarce/non-renewable natural resources (rare-earth elements, metals).
- High carbon footprint/greenhouse gas emissions from frequent manufacturing and transportation of replacement devices.
- High energy and water consumption required to refine raw materials for new electronics.
PastPaper.question 6 · Short Answer
2 PastPaper.marks
State two benefits of applying the data minimization principle when designing a new mobile fitness application.
PastPaper.showAnswers

PastPaper.workedSolution

Applying the data minimization principle—which dictates that an organization should only collect, process, and store the minimum personal data necessary to fulfill its specific purpose—offers key advantages:
1. **Reduced security risk**: If a data breach or cyberattack occurs, the volume of exposed sensitive user data (such as health metrics or location history) is restricted, limiting financial, legal, and reputational damage.
2. **Regulatory compliance and trust**: Designing with data minimization aligns with strict international laws (e.g., GDPR, CCPA), facilitating compliance and building user trust by demonstrating respect for user privacy.

PastPaper.markingScheme

Award [1 mark] for each distinct benefit of applying the data minimization principle, up to a maximum of [2 marks].

Suitable responses include:
- Lower risk of severe data breaches / smaller attack surface.
- Improved alignment and compliance with data protection laws (e.g., GDPR, CCPA).
- Enhanced user trust and brand reputation for prioritizing privacy.
- Reduced operational costs associated with storing, managing, and securing large volumes of unnecessary data.
PastPaper.question 7 · Medium Response
2 PastPaper.marks
Explain how a government-imposed internet shutdown during a national crisis can restrict the digital human rights of citizens.
PastPaper.showAnswers

PastPaper.workedSolution

A government-imposed internet shutdown completely blocks access to digital communication channels. This restricts citizens' right to freedom of expression and access to information (often aligned with Article 19 of the UDHR), as they cannot access news, share updates, or communicate with emergency services or family members during a crisis.

PastPaper.markingScheme

Award [1] mark for identifying a digital human right (e.g., freedom of expression, access to information) that is impacted. Award [1] mark for explaining how the shutdown specifically prevents citizens from exercising that right (e.g., unable to access news or communicate with family).
PastPaper.question 8 · Medium Response
2 PastPaper.marks
State two reasons why the enforcement of the 'Right to be Forgotten' (data erasure) presents challenges for global search engine companies operating across national borders.
PastPaper.showAnswers

PastPaper.workedSolution

First, global search engines face jurisdictional conflicts, where compliance with erasure requests in one region (e.g., Europe) conflicts with free speech protections in another (e.g., United States). Second, technical complexity makes indexing systems highly distributed; scrubbing all localized versions, mirrors, and cache nodes globally while keeping the search index coherent is difficult.

PastPaper.markingScheme

Award [1] mark for identifying jurisdictional conflict / differing international legal frameworks. Award [1] mark for identifying a technical hurdle / data distribution complexity, up to a maximum of [2] marks.
PastPaper.question 9 · Medium Response
2 PastPaper.marks
Describe how 'representation bias' in a training dataset can lead to discriminatory outcomes in an automated recruitment system.
PastPaper.showAnswers

PastPaper.workedSolution

Representation bias occurs when certain groups are underrepresented or overrepresented in the training data (for example, historical datasets containing mostly male profiles for technical roles). When an AI recruitment model is trained on this data, it associates dominant-group characteristics with job success, leading to systemic bias and discriminatory lower scores for applicants from underrepresented groups.

PastPaper.markingScheme

Award [1] mark for explaining how representation bias is established in the historical dataset. Award [1] mark for explaining how the machine learning system applies this bias to produce discriminatory outputs against minority candidates.
PastPaper.question 10 · Medium Response
2 PastPaper.marks
Distinguish between a 'teleological' (utilitarian) approach and a 'deontological' (duty-based) approach to programming the collision-avoidance system of an autonomous vehicle.
PastPaper.showAnswers

PastPaper.workedSolution

A teleological (utilitarian) algorithm determines the course of action that minimizes total harm or casualties, evaluating the ethical correctness based solely on the final outcome. In contrast, a deontological approach programs the vehicle to follow pre-defined ethical duties or rules (such as 'do not actively harm an innocent bystander' or 'maintain lane priority'), prioritising adherence to moral duties over the optimization of outcomes.

PastPaper.markingScheme

Award [1] mark for a clear explanation of the teleological approach focusing on outcomes/minimizing harm. Award [1] mark for a clear explanation of the deontological approach focusing on adherence to rules or duties.
PastPaper.question 11 · Medium Response
2 PastPaper.marks
Distinguish between an 'epistemic bubble' and an 'echo chamber' in the context of online information dissemination.
PastPaper.showAnswers

PastPaper.workedSolution

An epistemic bubble occurs when individuals do not encounter contrary views, typically due to algorithmic filtering, social selection, or passive exclusion (omission of voices). An echo chamber, however, is a structure where outside voices are actively discredited, mocked, or untrusted, meaning even if alternative information penetrates the group, it is systematically rejected.

PastPaper.markingScheme

Award [1] mark for defining an epistemic bubble as an informational network of omission/accidental exclusion. Award [1] mark for defining an echo chamber as an informational network of active exclusion and systematic discrediting of external sources.
PastPaper.question 12 · Medium Response
2 PastPaper.marks
Explain one environmental consequence of planned obsolescence in consumer electronics, and suggest one digital-society-driven policy measure to mitigate this issue.
PastPaper.showAnswers

PastPaper.workedSolution

Planned obsolescence encourages rapid replacement cycles, which leads to a massive accumulation of e-waste containing toxic heavy metals that pollute soil and water systems. A viable digital-society policy to mitigate this is the introduction of 'Right to Repair' legislation, which mandates that manufacturers provide software updates, repair documentation, and replacement parts for a prolonged period, extending device lifespans.

PastPaper.markingScheme

Award [1] mark for identifying and explaining a valid environmental consequence (e.g., e-waste, resource depletion, manufacturing emissions). Award [1] mark for proposing a valid regulatory or policy intervention (e.g., Right to Repair, mandatory recycling targets, software support extensions).
PastPaper.question 13 · Extended Response
8 PastPaper.marks
Multinational corporations are increasingly adopting artificial intelligence (AI) video-interviewing platforms. These platforms analyze job candidates' facial expressions, body language, vocal tone, and vocabulary to predict performance and cultural fit. Evaluate the ethical and social impacts of utilizing these AI-driven video-interviewing platforms in recruitment, with focus on bias, equity, and human dignity.
PastPaper.showAnswers

PastPaper.workedSolution

An outstanding response should address the following areas: 1. Introduction: Define AI-driven video-interviewing and identify the tension between corporate efficiency and ethical concerns such as bias, equity, and human dignity. 2. Ethical and Social Opportunities (Benefits): Standardization of the initial screening phase, which can theoretically eliminate human unconscious bias (e.g., affinity bias or prejudice based on gender/race from resume names); capability to process massive volumes of applicants quickly, giving all applicants an automated 'first round' opportunity. 3. Ethical and Social Challenges (Risks): Algorithmic bias, as the training data for these AIs often consists of historically successful employees, thereby reinforcing systemic gender, racial, or socioeconomic disparities; inequity for neurodivergent individuals or candidates with physical disabilities, whose speech patterns, facial expressions, or eye contact may deviate from the AI's programmed 'ideal' norm; erosion of human dignity, where complex human potential is reduced to a proprietary score generated by a 'black box' algorithm with zero transparency. 4. Synthesis and Evaluation: To what extent can these tools be used ethically? Candidates should argue that while these tools optimize logistical bottlenecks, they must not replace human judgment entirely. Solutions include explainable AI, regular bias audits, and clear opt-out clauses for candidates who require accommodations.

PastPaper.markingScheme

Marking Criteria:
- Level 3 (7-8 marks): The response demonstrates an in-depth understanding of the ethical and social impacts of AI recruitment. The evaluation is balanced, well-developed, and explicitly links digital concepts (bias, equity, human dignity) to the scenario. Terminology is precise and appropriate.
- Level 2 (4-6 marks): The response explains some ethical or social impacts. There is an attempt at evaluation, but it may be unbalanced or lack depth in critical analysis. Some appropriate terminology is utilized.
- Level 1 (1-3 marks): The response is largely descriptive, identifying a few basic points or impacts without structured evaluation. Terminology is limited or missing.
- 0 marks: No rewardable content.
PastPaper.question 14 · Extended Response
8 PastPaper.marks
The concept of 'smart cities' promises to revolutionize urban living by deploying thousands of Internet of Things (IoT) sensors to optimize traffic, waste management, and energy grids. However, the hardware lifecycle of these sensors raises significant ecological concerns. Discuss the environmental opportunities and challenges associated with the widespread deployment of IoT devices in smart cities.
PastPaper.showAnswers

PastPaper.workedSolution

An outstanding response should cover the following points: 1. Introduction: Define IoT in the context of smart cities and highlight the dual nature of its ecological footprint (resource saver vs. resource consumer). 2. Environmental Opportunities: Optimized energy grids where smart streetlights dim when empty, reducing municipal power usage; real-time water leak detection in aging pipe networks, conserving vital freshwater resources; dynamic traffic routing that reduces idle times for vehicles, leading to lower greenhouse gas emissions; optimized waste collection pathways, minimizing fuel consumption of city service fleets. 3. Environmental Challenges: Resource extraction, as millions of sensors require rare earth metals, driving destructive mining practices; energy footprint of the back-end infrastructure, where the continuous flow of real-time sensor data demands vast amounts of energy to run data centers; electronic waste (e-waste), since many distributed sensors are cheap, non-biodegradable, difficult to retrieve, and rely on chemical batteries that can leach toxic substances into the local soil and water tables upon disposal. 4. Conclusion/Synthesis: Balancing these factors requires cities to mandate circular economy principles, such as choosing modular, biodegradable, or solar-powered sensors, alongside green hosting for cloud infrastructure.

PastPaper.markingScheme

Marking Criteria:
- Level 3 (7-8 marks): The response shows a deep understanding of both the environmental benefits and ecological costs of IoT devices. The discussion is highly structured, balanced, and contains a well-synthesized conclusion using appropriate environmental terminology (e-waste, carbon footprint, circular economy).
- Level 2 (4-6 marks): The response discusses both opportunities and challenges, but may be unbalanced (e.g., focusing mostly on benefits with superficial coverage of e-waste/mining) or lack depth in environmental science connections. Some key terminology is used.
- Level 1 (1-3 marks): The response is descriptive, listing basic pros or cons of smart cities with limited focus on the technical or physical environment. Terminology is minimal.
- 0 marks: No rewardable content.

Paper 1 Section B

Answer one extended response question out of two. Each question is worth 12 marks and focuses on global challenges or interventions.
1 PastPaper.question · 12 PastPaper.marks
PastPaper.question 1 · Extended Response
12 PastPaper.marks
In many urban areas, governments are integrating artificial intelligence-driven facial recognition technology (FRT) into public surveillance networks to monitor crowds in real-time. While states justify this intervention as necessary for national security and crime prevention, civil society organizations argue it disproportionately infringes upon fundamental human rights.

Evaluate the effectiveness of different governance interventions (such as international human rights treaties, regional regulations, or complete bans) in balancing public safety and the protection of human rights in the context of state-deployed facial recognition technology.
PastPaper.showAnswers

PastPaper.workedSolution

### Key Arguments & Perspectives to Include:

1. **The Challenge: State Security vs. Human Rights**
* **State Perspective:** Proponents argue that real-time FRT is a vital tool for public safety, enabling immediate identification of suspects, preventing terrorist threats, and finding missing persons in crowded spaces.
* **Human Rights Perspective:** Opponents highlight severe risks to the UN Universal Declaration of Human Rights: Article 12 (Right to Privacy - mass surveillance), Article 19 (Freedom of Expression), and Article 20 (Freedom of Peaceful Assembly), noting a 'chilling effect' where citizens avoid public protests out of fear of state identification.
* **Bias and Discrimination:** Algorithmic bias in FRT (higher error rates for women and minority ethnic groups) leads to discriminatory policing practices.

2. **Evaluation of Governance Interventions:**
* **Complete Bans/Moratoriums:**
* *Effectiveness:* Highly protective of civil liberties. Prevents any risk of abuse. Adopted by cities like San Francisco and urged by the UN Human Rights Commissioner.
* *Limitations:* Can be seen as overly restrictive, preventing the legitimate use of technology to stop imminent physical threats (e.g., active shooters).
* **Regional Legislation (e.g., The EU AI Act):**
* *Effectiveness:* Uses a risk-based approach. Classifies real-time biometric identification in public spaces as an 'unacceptable risk' but allows very narrow, judicially authorized exceptions (e.g., kidnapping, targeted terror threats).
* *Limitations:* Enforcement can be complex, and exceptions may be exploited via 'mission creep' by state security apparatuses.
* **International Human Rights Treaties & Oversight:**
* *Effectiveness:* Sets global ethical baselines and moral pressure on nations.
* *Limitations:* 'Soft law' frameworks lack binding enforcement mechanisms, especially over sovereign superpowers that prioritize national security over international treaties.

3. **Integration of Digital Society Concepts:**
* **Power:** How FRT shifts the balance of power decisively toward the state, away from the citizen.
* **Values and Ethics:** The trade-off between the collective value of security and the individual value of liberty.

PastPaper.markingScheme

### Level Descriptors (Total 12 Marks):

* **Level 4 (10–12 marks):**
* Comprehensive, balanced evaluation of different governance interventions.
* Consistently integrates relevant Digital Society concepts (Power, Values and Ethics, Change).
* Uses specific, accurate real-world examples (e.g., EU AI Act, municipal bans, specific human rights frameworks) to support the analysis.
* A well-structured argument that leads to a logical, critical conclusion.

* **Level 3 (7–9 marks):**
* Good understanding of the tension between state security and human rights.
* Analyzes more than one governance intervention, but may focus heavily on one side.
* Includes relevant real-world examples, though some may lack depth.
* Clear structure with a reasoned conclusion.

* **Level 2 (4–6 marks):**
* Adequate understanding of the issues surrounding FRT and human rights.
* Description of governance interventions is present but lacks analytical depth.
* Examples are generic or hypothetical.
* Arguments may be one-sided or lack clear structure.

* **Level 1 (1–3 marks):**
* Minimal understanding of the topic.
* High reliance on descriptive narrative with little to no analytical focus on 'governance' or 'interventions'.
* No relevant real-world examples used.

Paper 3 Section A

Answer all questions. Refer to the provided source booklet, pre-released statement, and your own research.
6 PastPaper.question · 30 PastPaper.marks
PastPaper.question 1 · Short Answer
2 PastPaper.marks
Based on your study of digital governance and human rights, state two challenges that independent auditors face when attempting to assess proprietary algorithms for human rights compliance.
PastPaper.showAnswers

PastPaper.workedSolution

Independent auditing of algorithmic systems is an essential tool for human rights governance. However, auditors face significant barriers. First, companies often protect their proprietary software as trade secrets, denying auditors the access required to fully examine the system's inner workings. Second, translating abstract human rights concepts into mathematical code is challenging, especially because different mathematical models of 'fairness' often conflict with one another.

PastPaper.markingScheme

Award 1 mark for each valid challenge identified, up to a maximum of 2 marks. Suitable responses include: Intellectual property protections/trade secrets restricting access to code; difficulty accessing proprietary training data due to privacy laws; the 'black-box' nature of deep learning algorithms where decisions cannot easily be traced; and lack of standardized global frameworks for measuring algorithmic human rights impact.
PastPaper.question 2 · Short Answer
2 PastPaper.marks
Based on your study of sustainable development, state two environmental concerns associated with the widespread deployment of physical digital infrastructure, such as IoT sensors in smart cities.
PastPaper.showAnswers

PastPaper.workedSolution

While IoT networks can optimize resource management, their physical implementation has a significant environmental footprint. Millions of distributed sensors require rare earth metals and lithium batteries, creating a massive e-waste management challenge. Furthermore, the relentless stream of data generated requires continuous network transmission and computing power at data centers, escalating electricity demand and carbon emissions if fossil fuels are used.

PastPaper.markingScheme

Award 1 mark for each valid environmental concern identified, up to a maximum of 2 marks. Suitable responses include: E-waste generation from discarded hardware/batteries; energy usage from data transmission/processing centers; environmental degradation from raw resource extraction (such as lithium or cobalt for batteries); and chemical pollution from improper disposal of hazardous device components.
PastPaper.question 3 · Medium Response
3 PastPaper.marks
With reference to global initiatives for digital identity systems, outline three distinct barriers that can cause digital exclusion for marginalized communities during registration.
PastPaper.showAnswers

PastPaper.workedSolution

To implement a successful digital identity system, governments must address barriers to entry. These include:
1. Document barriers: Marginalized groups often lack baseline documents like birth certificates, preventing them from starting the registration process.
2. Geographic/Infrastructure barriers: Remote populations may face high travel costs and lack of internet infrastructure to reach registration stations.
3. Accessibility barriers: Biometric capture devices (iris or fingerprint scanners) may fail for manual laborers with worn fingerprints or individuals with physical disabilities.

PastPaper.markingScheme

Award [1] mark for each valid, distinct barrier outlined, up to a maximum of [3] marks.
- Accept answers referring to baseline documentation issues, physical infrastructure/location access, technological accessibility (biometrics), or lack of digital literacy/awareness.
- Do not award marks for repetitive or overly vague answers (e.g., 'they do not have technology' without elaboration).
PastPaper.question 4 · Medium Response
3 PastPaper.marks
With reference to the deployment of AI-driven digital health interventions, explain three ethical challenges developers face when designing conversational agents (chatbots) for mental well-being support.
PastPaper.showAnswers

PastPaper.workedSolution

Conversational agents in healthcare raise significant ethical concerns:
1. Privacy & Consent: Users disclose highly sensitive personal information. Developers must ensure secure encryption and prevent third-party monetization or unauthorized data access.
2. Algorithmic bias and safety: AI models trained on limited datasets may misunderstand non-Western idioms of distress, leading to inadequate support or incorrect risk triage.
3. Crisis escalation & Responsibility: If a user expresses suicidal ideation, a chatbot cannot feel empathy or take physical action. Clear protocols must exist to instantly hand off the interaction to human professionals, establishing clear lines of accountability.

PastPaper.markingScheme

Award [1] mark for each clearly explained ethical challenge, up to a maximum of [3] marks.
- Accept points discussing: Data privacy/consent, algorithmic bias/safety/misdiagnosis, user manipulation/dependency, or crisis response/liability/accountability.
- To achieve full marks, the explanation must link the ethical challenge directly to the context of mental health support or conversational agents.
PastPaper.question 5 · Extended Response
8 PastPaper.marks
The Aegis Coalition, a regional alliance of nations, has proposed a new regulatory framework mandating that all public-sector artificial intelligence (AI) systems must undergo independent, third-party Human Rights Impact Assessments (HRIAs) before they can be deployed.

Evaluate the effectiveness of mandating independent third-party HRIAs as a governance intervention to protect citizens' digital rights in this context.
PastPaper.showAnswers

PastPaper.workedSolution

Introduction:


Mandating independent third-party Human Rights Impact Assessments (HRIAs) is a proactive governance intervention designed to ensure that public-sector AI systems (e.g., predictive policing, automated welfare distribution, facial recognition) do not violate fundamental rights like privacy, non-discrimination, and due process.



Strengths/Benefits of the Intervention:



  • Objectivity and Trust: Utilizing external, third-party auditors reduces the risk of conflict of interest, ensuring that government departments do not simply self-approve biased or invasive systems. This fosters public trust in civic digital infrastructure.

  • Proactive Risk Mitigation: HRIAs force developers and public bodies to identify and mitigate potential algorithmic bias and human rights risks before the system is deployed, rather than reacting after harm has occurred.

  • Standardization: A regional framework establishes clear, shared ethical baselines across different public entities, aligning digital systems with international human rights standards (such as the Universal Declaration of Human Rights).



Limitations/Challenges of the Intervention:



  • The 'Moving Target' Problem: Public-sector AI systems, particularly those utilizing machine learning, adapt and evolve over time as they ingest new data. A static pre-deployment HRIA may fail to capture downstream drifts in algorithmic behavior.

  • Enforcement and Compliance 'Checklist' Culture: There is a risk that HRIAs turn into a superficial 'box-ticking' exercise (ethics washing), where agencies comply with the letter of the assessment without genuinely integrating human-centric values into their day-to-day operations.

  • Resource Constraints and Bottlenecks: Third-party audits require highly specialized legal and technical experts. This can introduce significant bureaucratic delays and high financial costs, potentially slowing down the delivery of beneficial public services.

  • Defining Metrics: Translating qualitative human rights concepts (e.g., 'fairness' or 'dignity') into quantifiable parameters for AI evaluation remains technically and philosophically challenging.



Conclusion/Synthesis:


While third-party HRIAs are a highly valuable mechanism for establishing transparency and systemic accountability, their effectiveness is contingent on continuous post-deployment monitoring and the establishment of robust, legally binding penalties for non-compliance. Without iterative oversight, HRIAs risk becoming expensive administrative formalities rather than true safeguards for digital rights.

PastPaper.markingScheme

Level descriptors (8 Marks total):



7–8 Marks:



  • The response shows a highly detailed and accurate understanding of the role of Human Rights Impact Assessments (HRIAs) as a digital governance mechanism.

  • The evaluation is balanced, explicitly contrasting multiple strengths (e.g., objective oversight, proactive bias mitigation) against realistic limitations (e.g., adaptive AI drift, compliance costs, ethics washing).

  • The argument is well-structured and leads to a reasoned, synthesized conclusion that directly answers the prompt. Relevant digital society concepts and terminology are applied accurately.



5–6 Marks:



  • The response demonstrates a good understanding of HRIAs and public-sector AI systems.

  • There is a clear attempt at evaluation, though it may favor either strengths or weaknesses somewhat unevenly.

  • The connection to citizens' digital rights (e.g., privacy, non-discrimination) is made, but some arguments may lack depth or technical precision.



3–4 Marks:



  • The response is largely descriptive, explaining what HRIAs or AI bias are without fully evaluating the *intervention* itself.

  • Points raised are general, with limited reference to specific digital rights or the implications of third-party public-sector governance.

  • The structure may be weak, and any conclusion is superficial or unsupported.



1–2 Marks:



  • The response is minimal or contains major misunderstandings about AI, human rights, or governance.

  • Few or no relevant digital society terms are used.

PastPaper.question 6 · Recommendation
12 PastPaper.marks
The municipal government of Neo-Veridia is facing a crucial decision regarding public safety and digital rights. They are considering two alternative paths:

* **Option A**: Implement 'BioWatch', a centralized, real-time facial recognition and biometric surveillance system integrated with existing CCTV networks to proactively detect and deter criminal activity.
* **Option B**: Invest in 'CivicSafe', a decentralized, community-led digital safety platform that allows citizens to report safety concerns anonymously, share localized alerts, and coordinate voluntary community watch programs without active government biometric tracking.

Recommend which option the municipal government of Neo-Veridia should adopt. Justify your recommendation by evaluating both options, referring to governance, human rights, and the ethical implications of digital surveillance.
PastPaper.showAnswers

PastPaper.workedSolution

### Analysis of Option A (BioWatch)
* **Strengths:**
* Highly efficient real-time monitoring and rapid law enforcement deployment.
* Strong deterrent effect against public crimes and acts of violence.
* Automation reduces the need for continuous manual surveillance of video feeds.
* **Weaknesses and Human Rights Concerns:**
* Severely compromises the fundamental right to privacy (Article 12 of the Universal Declaration of Human Rights) and creates a constant 'chilling effect' on freedom of expression and peaceful assembly.
* High risk of algorithmic bias, as facial recognition models historically demonstrate higher error rates for women and racial minorities, potentially leading to systemic discrimination.
* High susceptibility to centralized data breaches and mission creep (e.g., using the data for purposes beyond crime prevention).

### Analysis of Option B (CivicSafe)
* **Strengths:**
* Preserves civilian privacy and avoids the risks associated with mass biometric state surveillance.
* Empowers community agency, collaboration, and trust between citizens and local authorities.
* Lower infrastructure costs and avoids the ethical and technical pitfalls of biometric classification systems.
* **Weaknesses and Technical Challenges:**
* Relies heavily on active citizen engagement and assumes widespread digital literacy (potentially marginalizing those on the wrong side of the digital divide).
* Risk of citizen-reported misinformation, bias in crowd-sourced alerts, or vigilante behaviors.
* Less effective for instantaneous detection of high-speed, severe threats compared to automated video analysis.

### Recommended Course of Action and Justification
The municipal government should adopt **Option B (CivicSafe)**. While public safety is a legitimate state interest, the deployment of a centralized biometric system (Option A) poses disproportionate risks to fundamental human rights, civil liberties, and equality.

To address the limitations of Option B:
1. **Governance Framework:** Establish clear, independent oversight boards to moderate crowd-sourced safety reports, preventing vigilante profiling and misinformation.
2. **Inclusivity Measures:** Provide analog alternatives (such as physical community hubs or free telephone hotlines) to bridge the digital divide for older or lower-income residents.
3. **Hybrid Security:** Combine CivicSafe with traditional, non-biometric community policing methods to maintain safety without resorting to mass-surveillance technologies.

PastPaper.markingScheme

**[10–12 Marks]**
* The response presents a highly focused, balanced, and persuasive recommendation that directly addresses the tension between public safety and human rights.
* Both options are evaluated in detail, demonstrating an excellent understanding of digital concepts (e.g., algorithmic bias, biometric data, decentralized networks) and contexts (governance, ethics, human rights).
* The recommendation is fully justified with strong analytical reasoning and proposes realistic mitigation strategies for the chosen option's weaknesses.

**[7–9 Marks]**
* The response presents a clear recommendation and evaluates both options, but the analysis may favor one option without fully exploring the counterarguments or limitations of both.
* Explains relevant digital concepts and human rights frameworks, but some connections may lack depth.
* The justification is logical but may miss some critical trade-offs or fail to propose effective mitigation strategies.

**[4–6 Marks]**
* The response provides a recommendation but lacks a balanced evaluation of both options.
* The understanding of digital society concepts is basic, or the discussion of human rights and governance is superficial.
* The justification is weak or relies on unsupported assertions.

**[1–3 Marks]**
* The response is highly descriptive, fails to make a clear recommendation, or shows minimal understanding of the digital scenario and associated ethical issues.

PastPaper.sampleCTATitle

PastPaper.sampleCTADescription

PastPaper.ctaPrimaryPastPaper.ctaSecondary

PastPaper.sampleStickyMessage

PastPaper.stickyCtaText