Computing decisions affect different people in different ways. Build an argument that explains a mechanism, considers a counterargument and reaches a justified judgement.
Content owner: Michael Print · Written for A-Level learners · Checked against official specifications
The idea to start with
An ethical evaluation asks who benefits, who carries the cost, what evidence supports the claim and how the design could reduce harm. Legal compliance and ethical acceptability are related questions, but they are not interchangeable.
A strong paragraph follows decision → mechanism → effect → qualification. For example, an automated booking system can reduce clerical work, but people without reliable internet may need a telephone alternative.
OCR H446 · 1.5.2 (all nine listed issue areas)
Before you start
Useful foundations
Distinguish a technical capability from its social consequences.
Identify stakeholders: people or organisations affected by a decision.
By the end, you should be able to
Evaluate all nine OCR issue areas in context.
Connect benefits, harms and mitigations to specific stakeholders.
Write a supported conclusion rather than a list of opinions.
Workforce, automated decisions and AI
Automation may remove repetitive tasks, change roles and create demand for maintenance or technical skills. The benefits depend on training, transition time and who receives the productivity gains. Avoid claiming that every automated task removes an entire job.
AI-supported decisions can process large datasets consistently, yet errors, biased training data or unsuitable assumptions may affect groups differently. For high-impact decisions, ask whether people can understand the relevant reasons, challenge an error and reach a responsible human.
A human review is only useful if the reviewer has time and authority to act.
Environmental effects over the whole life cycle
Repairability, useful life, electricity consumption and recycling all matter. A remote meeting may save travel, but its equipment and network still consume resources.
Compare alternatives using the same workload and time period. Replacing working laptops with lower-power models is not automatically the best environmental decision: the manufacturing cost and remaining useful life must also be weighed. Explain what evidence would be needed rather than inventing an emissions figure.
Assess the whole device life cycle
1
Manufacture
Resources are used before the device begins work.
2
Transport
Getting devices to their users adds impacts.
3
Operation
Compare electricity consumption over the same workload and period.
4
Disposal or reuse
Account for remaining useful life, repairability, recycling and waste.
Censorship, monitoring and personal information
Content moderation may reduce harassment or illegal material while mistaken filtering can suppress legitimate speech. A rule needs an understandable boundary and a way to review errors. Discuss the effect on participants rather than simply labelling all filtering good or bad.
Monitoring may support safety or detect misuse; continuous collection can also reveal private habits and discourage legitimate behaviour. Collecting a location for one journey does not justify storing a person's movements indefinitely. Consider purpose, proportionality, access, retention and the ability to challenge decisions.
Piracy and offensive communications
Unauthorised copying can deprive creators of licensing income and distribute versions without security support. Access and affordability may explain a user's motivation, but do not by themselves settle legality or ethical responsibility. Compare licensed alternatives, open resources and sustainable payment models.
Offensive online communications can cause harm at scale and persist through reposting. Reporting tools, moderation and user controls may help, though false reports and overbroad removal also affect users.
Explain an actual harm or trade-off in the given context; avoid substituting a personal dislike for an argument.
Inclusive layout and colour
An interface should match users' needs: readable type, clear grouping, meaningful labels and consistent navigation reduce errors. Colour alone may exclude someone who cannot distinguish the chosen colours; labels or symbols provide another cue.
Character sets and cultural assumptions
A character set and encoding affect which languages and symbols can be stored and displayed. ASCII alone cannot represent the range of names required by a multilingual service.
Unicode provides a much larger repertoire, while the chosen encoding and fonts must support correct storage and display. A cultural assumption in date formats or name fields can be as important as visual styling.
Make a judgement with conditions
Choose the most important effects for the situation. A recommendation such as “use the system with an appeal process and an offline route” follows from evidence when it addresses identified errors and exclusion.
Build an evaluation that earns its conclusion
1
Decision
Identify what the proposed system will do.
2
Mechanism
Explain how that choice causes a benefit or harm.
3
Effect
Connect the consequence to a specific stakeholder.
4
Qualification
Compare an alternative, propose a safeguard and identify uncertainty.
5
Judgement
Reach a supported recommendation with practical conditions. Generic “advantages and disadvantages” do not establish it.
Worked example
A library introduces automated access
The proposal uses a smartphone code to unlock the building and logs each visit. Staff can focus on support rather than routine entry checks.
Some visitors do not own smartphones. Offer a staffed or physical-card route; otherwise the mechanism of access excludes precisely those users.
Visit logs may help investigate incidents, but indefinite retention could reveal reading habits. Restrict access and define a justified retention period.
An automated rejection must offer a clear explanation and reachable staff support. Replacing a staffed entrance without a workable exception route shifts the burden onto vulnerable users.
Judgement: trial the system with alternative access and limited logging, measuring waiting time, failed entries and complaints before widening use. This balances the claimed efficiency against evidenced exclusion and privacy risks.
Worked example
A discussion service filters and profiles its users
A service removes messages matching a banned-word list and monitors clicks to predict users' interests. Blocking a word can remove both an abusive message and a legitimate discussion of that abuse: the same mechanism has different effects by context.
Human review and an appeal route can reduce mistaken censorship. State who can review the context and reverse an error; a vague promise of moderation is insufficient.
Combining repeated clicks and visit times may reveal private interests. Explain what predictions the profiling produces and whether the information is necessary for the service's stated purpose.
A defensible design reduces monitoring to necessary data, explains the use, limits access/retention and evaluates mistaken removals. These are conditions for the recommendation, rather than proof that every remaining risk disappears.
Worked example
A club shares a paid game and abusive messages
A club member uploads an unauthorised copy of a paid game to save other members money. Explain affordability as a motivation, then consider the creator's income, licence conditions and the risks of unsupported copies.
Compare a licensed shared-use arrangement, a discount or an open-source alternative. A better-access argument does not automatically remove the creator's interests.
Members then repost humiliating messages about a player. The harms may persist through copying and affect participation, even after the first post is removed.
Recommend a clear reporting/moderation route and proportionate review, while protecting legitimate criticism from arbitrary removal. Tie the judgement to both the copying decision and the communications harm.
Original A-Level practice
6 original questions total 32 marks. Attempt each before opening the independently written indicative marking guidance.
Question 1
4 marks
A warehouse automates item picking. Explain one benefit and one workforce concern, developing both in context. [4]
Show solution and marking guidance+
Indicative answer
For example, routine picking may become faster or more consistent, reducing order delays (1+1).
Some staff roles may shrink or change; workers need realistic retraining or redeployment rather than an unexplained promise of new jobs (1+1).
Question 2
4 marks
A booking form displays errors only in red and accepts names only from ASCII. Explain two inclusion problems and suitable improvements. [4]
Show solution and marking guidance+
Indicative answer
Some users cannot distinguish the red cue; add explicit error text or symbols (1+1).
ASCII excludes many names/scripts; use Unicode-capable storage/display and appropriate validation (1+1).
Question 3
4 marks
Give two reasons why replacing every working computer with a lower-power model may not reduce total environmental impact. [4]
Show solution and marking guidance+
Indicative answer
Manufacture/transport of replacements uses resources; account for that impact against operational savings over a stated period (1+1).
Early disposal creates waste and loses remaining useful life; compare repair/reuse with replacement (1+1).
Question 4
6 marks
Evaluate a school proposal to use an AI system to flag students for sanctions from message histories. Reach a justified conclusion. [6]
Show solution and marking guidance+
Indicative answer
Develop a context-specific benefit, such as finding threatening behaviour sooner, including how it occurs (up to 2).
Develop a risk such as false positives, bias, privacy intrusion or chilling legitimate conversation (up to 2).
Give a justified judgement with a practical condition, such as minimised collection, informed policy and accountable review/appeal, linked to the competing effects (up to 2). A list of generic pros/cons is insufficient for all marks.
Question 5
8 marks
A discussion service automatically removes messages containing listed words and profiles users from every click. Evaluate censorship and personal-data concerns, giving a justified design recommendation. [8]
Show solution and marking guidance+
Indicative answer
Develop how filtering may reduce harmful messages or exposure (up to 2).
Explain context loss and mistaken censorship, including effects on legitimate speech (up to 2).
Explain how behavioural monitoring/profiling can infer private interests and why collection scope matters (up to 2).
Recommend a justified combination such as contextual appeal/review and minimised, explained data use, linked to the competing effects (up to 2).
Question 6
6 marks
Evaluate a club member's proposal to distribute an unauthorised game copy and to repost humiliating messages about another player. Give a reasoned alternative. [6]
Show solution and marking guidance+
Indicative answer
Develop the copying issue, weighing access/affordability against creator/licence interests or unsupported-copy risks (up to 2).
Develop the harm from offensive communications and persistence/reposting, linked to the affected player (up to 2).
Give justified alternatives such as a permitted licence/open resource plus proportionate reporting and review, addressing both issues (up to 2).
Specification and references
This guide addresses OCR H446 1.5.2 (all nine listed issue areas). Check your examination year and the complete specification for the assessment scope.
These are independently written explanations and practice questions. CompSciTutoring.co.uk is not affiliated with or endorsed by an examination board. The marking guidance is indicative; always check the syllabus for your examination year.