Paper 1 Ethics Issues Answers
These answers correspond to Paper 1 Ethics Issues Drills.
Answer 1: PDPA Obligation
One likely breached obligation is consent or purpose limitation. The parent supplied the email address for lesson invoices, not partner advertising. The centre reused or disclosed the email address for a new marketing purpose without telling the parent and obtaining agreement for that purpose.
Mark points:
- names a relevant obligation such as consent, notification, or purpose limitation;
- links the original purpose to lesson invoices;
- explains that partner advertising is a different undisclosed purpose.
Common weak answer:
- saying only “PDPA is broken” without naming or applying an obligation.
Answer 2: Consent
The statement is weak because it does not tell users the specific purposes for using the data. It is also too broad: users cannot make an informed choice about use of their phone number or dietary requirements if the app does not explain what each item will be used for.
Mark points:
- consent is not informed because the purpose is unclear;
- consent is not purpose-specific because “may be used” is too broad.
Common weak answer:
- saying consent is valid whenever a box is ticked. The user must understand what they are agreeing to.
Answer 3: Data Protection
Two suitable safeguards:
- restrict database access to authorised clinic staff using accounts and permissions;
- encrypt the database or backups so stolen files are not easily readable.
Other valid safeguards include audit logs, secure backups, staff training, and strong authentication.
Mark points:
- gives one relevant safeguard;
- gives a second relevant safeguard.
Common weak answer:
- saying “keep it safe” without a concrete technical or organisational control.
Answer 4: Professional Ethics
Three responsible actions:
- report the exposed CSV promptly to the appropriate supervisor/security contact;
- avoid downloading, sharing, or using the student data except as needed to report the flaw;
- help remove or restrict access to the file and preserve enough evidence for investigation.
Mark points:
- reports responsibly;
- avoids exploiting or spreading the data;
- protects affected users by supporting containment or investigation.
Common weak answer:
- posting the exposed URL publicly to prove the problem exists. That increases harm.
Answer 5: Bias
The model learns patterns from the historical data. If Group B was approved less often in the training records, the model may treat group-related patterns as signals for rejection. This can reproduce or amplify historical unfairness, causing similar current applicants from Group B to be recommended less often.
Mark points:
- identifies imbalance or bias in training data;
- explains that the model learns from historical patterns;
- links this to unfair lower recommendations for Group B.
Common weak answer:
- saying AI is always neutral because it uses mathematics. The data and design choices can encode bias.
Answer 6: Stakeholders
Three stakeholders include:
- students whose faces and attendance are recorded;
- parents who may receive attendance information or worry about privacy;
- school staff or administrators who operate and rely on the system.
Other valid stakeholders include system vendors, security staff, visitors, and regulators.
Mark points:
- identifies one affected stakeholder group;
- identifies a second affected stakeholder group;
- identifies a third affected stakeholder group.
Common weak answer:
- listing only “the school” three times in different words.
Answer 7: Social Impact
Positive impact: users may borrow books faster without queuing, especially during busy periods.
Negative impact: users who are elderly, have disabilities, or are unfamiliar with technology may find the service harder to use if human support is reduced.
Mark points:
- describes one positive impact;
- links the positive impact to the kiosk scenario;
- describes one negative impact;
- links the negative impact to the kiosk scenario.
Common weak answer:
- giving only business impacts such as lower staff cost when the question asks for social impact.
Answer 8: Economic Impact
Two possible economic impacts:
- the company may reduce costs or process claims faster because fewer staff hours are needed for simple claims;
- some clerical roles may be reduced or changed, while new roles may be needed for maintaining, auditing, or improving the system.
Mark points:
- explains cost/productivity impact;
- explains job displacement or job-role change;
- links both to automated claims processing.
Common weak answer:
- saying computing always creates more jobs overall. It can create some roles while reducing others.
Answer 9: Legal vs Ethical
Legal compliance means the company follows the law, such as keeping logs only for a permitted period. Ethical behaviour also considers fairness, transparency, and user expectations. Even if one-year retention is legal, using support logs to train a chatbot may be unethical if customers were told the data was only for support and were not informed about the new purpose.
Mark points:
- distinguishes law/compliance from broader ethical responsibility;
- applies the distinction to chatbot training beyond the original support purpose.
Common weak answer:
- saying legal and ethical always mean the same thing.
Answer 10: Mitigation
Four suitable controls:
- explain main rejection reasons to applicants where practical;
- provide an appeal or human review process;
- test approval rates for unfair bias across groups;
- keep audit logs of model inputs, outputs, and decisions.
Other valid controls include data minimisation, regular model review, transparency notices, and access control over applicant data.
Mark points:
- gives a transparency/explanation control;
- gives an appeal or human review control;
- gives a bias testing/monitoring control;
- gives an audit/security/data-governance control.
Common weak answer:
- saying “make the AI accurate” without stating how harm will be detected or reduced.