Jigsaw — 7 Ethics and Ownership (AS Level)

Jigsaw — 7 Ethics and Ownership

Syllabus content for 9618 Topic 7, rebuilt bullet by bullet, with every tested angle mapped onto it.

Legend

Already examined in 9618

Tested in a 9618 paper (2021 onwards). Latest question ID given.

9608 only (not yet in 9618)

Tested under the old 9608 syllabus, still inside the 9618 syllabus wording. Fair game — just untested in the current series. Latest 9608 question ID given.

Not yet tested — inference

In syllabus, not yet asked in either series (or only asked in a much narrower form). Justification given.

From the Save My Exams notes

Content the SME revision notes teach that no past question above covers. Weigh it against the mark schemes — SME sometimes goes beyond what Cambridge credits.

AI is new in 9618

The 9608 Legend’s AI section is empty — not one 9608 question. Every AI question in existence is 9618, 2021 onwards, and the sub-topic has appeared in every single series. There is no 9608 safety net here, so the !info inferences in 7.1.5 carry more weight than anywhere else in this note.

---

7.1.1 Show understanding of the need for and purpose of ethics as a computing professional

Describe the purpose of a code of conduct

To create a safe, respectable and professional working environment for all employees · to ensure employees understand the expectations and the consequences of their actions · to provide a clear outline of what is considered good practice · so there are clear rules to follow in certain situations · to protect the reputation of the organisation // maintain public trust. 9618_w25_qp_12_sc_2.b

Explain why it is important to join a professional ethical body

The body has ethical guidelines to follow … so clients/staff know the standards being applied … so the professional does not have to decide what is ethical, it is written down · clients know the member is reputable … there may be a test/entry requirement … giving recognition of skills and knowledge · the body provides help and support, e.g. legal advice · the body runs training courses … to keep skills up to date. 9618_w24_qp_13_sc_5.b

Describe the consequences of not joining a professional ethical body

No or limited access to legal advice, in case action is taken against them · fewer networking opportunities, so they could miss out on contacts/jobs · less access to training · no clear laid-out ethical guidelines, and no one to discuss potential ethical problems with, possibly leading to inappropriate or unethical actions, which might lead to legal proceedings. 9618_s25_qp_13_sc_3.d

Describe the benefits to an individual of joining a professional ethical body

9618_w21_qp_11_sc_4.a, 9618_w23_qp_13_sc_6.b

Explain the reasons why a company needs a professional code of conduct

To identify what employees can and cannot do · to identify the repercussions of employees performing activities they should not · to identify the company’s and the employee’s responsibilities · to identify the company’s values/missions · to identify what behaviour can be measured against · to identify how they will comply with legislation. 9618 asks about the code’s purpose in general; the company-side framing, and the compliance and measurement points, are 9608’s. 9608_s21_qp_13_sc_7.c

Explain why an employee is asked to sign a code of conduct before starting work

The employee is confirming they understand it · to make sure they know what behaviour is expected · to make sure they understand there may be consequences for some actions · to ensure all employees adhere to the same standards. 9608_s20_qp_12_sc_4.b

Describe what is meant by ethics

A system of moral principles that guide behaviour and decision making, based on philosophical or religious views, by example (respectful and considerate behaviour). The bare definition has never been asked in 9618, which always wraps ethics in a scenario. 9608_w17_qp_13_sc_6.a

The ACM/IEEE Software Engineering Code of Ethics — the eight principles. Name and describe them

Public — act consistently with the public interest · Client and Employer — act in the best interests of client and employer, consistent with the public interest · Product — ensure products and related modifications meet the highest professional standards · Judgement — maintain integrity and independence in professional judgement · Management — managers and leaders shall promote an ethical approach to software development and maintenance · Profession — advance the integrity and reputation of the profession · Colleagues — be fair to and supportive of colleagues · Self — participate in lifelong learning and promote an ethical approach to the practice of the profession. This framework was examined repeatedly in 9608 (named principles, matching incidents to principles, illustrating principles with examples) and has not appeared once in 9618, which names only BCS and IEEE as bodies. 9608_s19_qp_11_sc_4.a, 9608_s16_qp_12_sc_6

BCS and IEEE by name

The syllabus notes name BCS (British Computer Society) and IEEE (Institute of Electrical and Electronic Engineers) explicitly. Every 9618 question says only “a professional ethical body” — no question has ever required the names, what they stand for, or what each does. A 1–2 mark “name two professional ethical bodies” or “state what the BCS provides for its members” is free marks sitting unclaimed in the syllabus notes.

The benefits, inverted for the employer or the client

All four 9618 questions take the professional’s point of view (benefits to him, consequences for her). Why a client would prefer to hire a member of an ethical body, or why an employer would require it, is the same content from the other side and has not been asked.

What an ethical body actually is

An organisation that publishes a code of conduct which its members are expected to follow; membership signals a commitment to professional integrity and high standards. BCS — the professional body for IT and computing in the UK; promotes ethical behaviour, good practice and high standards; members follow a Code of Conduct. IEEE — a global organisation for engineers, developers and computer scientists; sets technical standards and publishes research and codes of ethics. This is the content for the !info gap above — and note the two are not interchangeable: BCS is national and professional-conduct focused, IEEE is global and standards-focused.

Why ethical issues in computing are hard

Technology moves faster than the law, there may be no single correct answer, and choices can have serious consequences for individuals and society. A framing no mark scheme gives, but a strong opening line for any “explain why this is an ethical issue” question.

The three key areas of computer ethics

Intellectual property rights (it is unethical to copy or use software, music or digital content without permission) · Privacy (users have the right to keep personal information safe; accessing someone’s data without permission is unethical and illegal) · Impact on society (job losses through automation, social media effects, digital addiction). A useful three-way structure for scenario answers, and it links this bullet to copyright below and to AI’s social impact.

Plagiarism and how to avoid it

The internet has made plagiarism easier — copying another person’s work and claiming it as your own. Using someone’s idea is not wrong, but credit must be given, through footnotes or references/citations. Anti-plagiarism software scans and detects copied content. Plagiarism appears in the 9618 mark schemes only as “students should not copy work that is not their own” (w25_qp_11_sc_8.b) — the citation mechanism and anti-plagiarism software are not credited anywhere.

---

7.1.2 Show understanding of the need to act ethically and the impact of acting ethically or unethically for a given situation

Explain why acting ethically towards colleagues and the public matters

Colleagues: treat them fairly and avoid discrimination · be prepared to accept and offer critique, so technical work improves · credit colleagues’ contributions, so they feel valued and respected · help and train colleagues. The public: maintain their health, safety and welfare, so as not to endanger anyone · be honest and realistic in claims about the software, so public trust is maintained · maintain the security of the public’s data. 9618_w24_qp_12_sc_5.a

Explain why it is important to act ethically towards a team

So team members feel valued · to get the best work out of the team · to enable them to work well together · to enable them to create the best product for the client. 9618_s21_qp_12_sc_2.a

Explain why someone acted unethically in a scenario (hiding a known error)

He did not act in the best interest of the product (it might fail) / the client (they are let down) / the profession (he lets it down) / the company (not correcting the error early leads to later problems). 9618_w21_qp_11_sc_4.c

Describe how someone can act ethically in a difficult situation (unfamiliar tool)

Tell the manager he has not used it and how he will get up to date · do his own research · explain that he needs additional training · ask to be booked on a training course · ask for a mentor / to shadow someone · practise at home before starting. 9618_w21_qp_11_sc_4.b.i

Give ethical considerations for a student using a personal computer on a school network

Consider privacy issues · do not make inappropriate use of social networking / do not engage in cyberbullying · do not copy work that is not their own · abide by the school’s rules for accessing the network · use the network only for legitimate purposes · do not hack or crack other computers. 9618_w25_qp_11_sc_8.b

Describe an ethical impact of AI in a given situation

9618_w25_qp_13_sc_7.a — see 7.1.5.

Tick ethical or unethical for a scenario, and justify your choice

The signature 9608 format, run across six separate scenarios (using a non-standard IDE; hiring your own daughter; dropping a concern after one raise; choosing a quick but inefficient algorithm; forcing a team to cancel holidays; discussing project concerns on a public forum). Crucially the mark schemes accept either verdict if the justification is sound — the marks are entirely in the reasoning, not the tick. 9618 always tells you the behaviour was unethical and asks why; it has never asked the candidate to make and defend the judgement. 9608_s21_qp_12_sc_6.a, 9608_s21_qp_11_sc_7.c

Classify incidents as ethical or unethical and map each to an ACM/IEEE principle

e.g. reporting a phishing email → ethical, Public · copying beta software and selling it → unethical · refusing to speak to a difficult colleague → unethical · reporting remaining bugs to the manager → ethical, Product. 9608_w17_qp_12_sc_5.a, 9608_w17_qp_12_sc_5.b

Explain how to act in the best interest of the client

Keep the client’s personal data private · involve the client in development / communicate with them · provide the solution the client asked for · keep the project on time and on budget · keep the client informed of any problems or delays. 9618 covers colleagues and the public (w24_qp_12_sc_5.a) but the client is only ever a mark point inside other answers. 9608_s19_qp_13_sc_2.d

Ethical actions on starting a new job, from three angles

What the new employee can do (read about the languages, visit the office beforehand, speak to the manager about concerns) · what the manager can do (prepare an induction, invite them in before starting, introduce them to the team, give them a mentor) · what colleagues can do. 9608_s20_qp_12_sc_4.a.ii

The candidate making the ethical judgement

This is the largest structural gap in Chapter 7. Across six 9608 scenarios candidates had to decide whether behaviour was ethical and defend it, with both verdicts creditable. Every 9618 question pre-loads the verdict. The syllabus wording — “the impact of acting ethically or unethically for a given situation” — supports the 9608 format exactly, and a scenario with a genuine two-sided answer (deadline vs quality, loyalty to employer vs public safety) is entirely available.

The impact of acting ethically, stated positively

9618 asks for the reasons why one should act ethically and the reasons why an action was unethical. A question asking for the consequences of acting ethically (reputation, trust, repeat business, staff retention) or the consequences of acting unethically (legal action, loss of clients, damage to the profession) has not been set, though s25_qp_13_sc_3.d is the nearest.

Whistleblowing / raising a concern

The 9608 Jason and Debbie scenarios both turn on whether and how to escalate a concern, and both are creditable either way. 9618’s w21_qp_11_sc_4.c touches it (not reporting an error) but never asks about the route — internally, to a manager, to a professional body, or publicly.

Worked ethical-conduct examples, ready to adapt

A project manager leading developers: treat all team members fairly, listen to their ideas, avoid favouritism (they feel valued) · respect work–life balance by avoiding unrealistic deadlines (better teamwork and better work) · encourage collaboration and clear communication (a better product for the client). An IT company owner: provide fair pay and ongoing training (staff feel valued, performance improves) · promote ethical behaviour with clients — do not overcharge or mislead (builds trust) · create a culture of respect even under pressure. The work–life-balance and fair-pay angles do not appear in any mark scheme here, but are exactly the “impact” half these questions ask for.

---

7.1.3 Show understanding of the need for copyright legislation

Describe what is meant by copyright

The formal and legal rights to ownership // intellectual property rights · protects against unauthorised reproduction of work · provides for a legal right of redress. 9618_s21_qp_11_sc_1.d

Describe copyright in a table alongside other terms

“Gives the holder the legal right of ownership of intellectual property.” 9618_w25_qp_12_sc_2.a

Explain the reasons why a programmer should copyright their program

To identify themselves as the owner/author // gain formal recognition of ownership · to allow legal consequences if anyone steals or copies it · to restrict competition. 9618_w24_qp_12_sc_5.b.ii

Identify ways to prevent illegal copies being installed

Encryption · assign a unique code / product key so it will not install without one · limit the number of times that version can be installed · set a time limit within which the software must be installed. 9608_w19_qp_12_sc_5.a

Identify a way to distribute software without the source code

Provide an .exe file // compile the source code // use a compiler. A neat cross-link to Chapter 5 (translators) and to the open-source/commercial distinction. 9608_w19_qp_12_sc_5.b

Why copyright legislation is needed — the societal argument

Every question is about what copyright is or why an individual should use it. The bullet says “the need for copyright legislation” — that without it creators cannot earn from their work, there is no incentive to create, and no legal route to redress. That argument has not been asked for directly in either series.

Copyright applied to something other than a program

s21_qp_11_sc_1.d uses an image and w24_qp_12_sc_5.b.ii a program. Music, video, text and databases are all covered by the same legislation, and a scenario involving downloading or sharing media is untested in 9618.

The Copyright, Designs and Patents Act — the named legislation

Protects the intellectual property of an individual or company; makes it illegal to copy, modify or distribute software or other intellectual property without permission. Copyright is applied automatically on creation of an original work and lasts, depending on the type of work, until roughly 25–70 years after the creator’s death. It is the owner’s responsibility to take action if they believe their work has been copied. Primary breaches: copying an original work · issuing a copy to the public · renting or lending a copy to the public · performing, showing or playing the work in public · making an adaptation. Secondary breaches: importing a copy · possessing or dealing with a copy · providing the means to make copies · permitting premises to be used for making copies · providing equipment for a performance of a copy. No 9618 mark scheme in this topic names the Act or these breaches; “automatically applied on creation” is the single most useful addition, because it is what makes the 9618 answer “formal and legal rights to ownership” true without registration.

Examiner note on naming the Act

If you name it, use the full name — the Copyright, Designs and Patents Act. “The Copyright Act” is not its name and a shortened version may cost the mark. And the standard structure for an applied answer is: the work was protected by the Act when it was created, therefore anyone wishing to distribute it must gain the owner’s permission, so posting it online would breach the Act.

Copyright and streaming

The Act covers video and audio, where peer-to-peer streaming or sharing prevents the copyright owner from receiving an income — e.g. downloading from a streaming service and sharing the files. The “prevents the creator earning their income” framing is the clearest answer to why the legislation is needed.

---

7.1.4 Show understanding of the different types of software licensing, and justify the use of a licence for a given situation

Describe each licence type in a table

Copyright — gives the holder the legal right of ownership of intellectual property · Open Source (Initiative) — software can be freely copied, distributed or adapted · Shareware — gives the user a trial period / limited functionality before a payment or sign-up may be required · Software licence — a legal agreement that defines the rights and terms of usage for the software. 9618_w25_qp_12_sc_2.a

Describe the features of a commercial software licence

The program is usually sold for a fee · users have no access to source code · it protects the intellectual rights of the programmer · users cannot legally alter or share the program · there are limitations on use, e.g. the number of installations · support and maintenance are usually provided (updates, bug fixes). 9618_s25_qp_12_sc_3.c.i

Describe what is meant by Free Software

The software is free of restrictions · it is not necessarily free of charge · the source code comes with the program, and a user can edit it to add functionality or suit their needs · users can share the software with others. 9618_w25_qp_12_sc_10.b

Identify two licences allowing others to edit and redistribute

Open Source (Initiative) and Free Software (Foundation). 9618_w25_qp_13_sc_8.c

Match a statement to its licence type

“A fee is charged, source code cannot be accessed, no free trial” → Commercial · “users can try before buying, may not access all features during the trial” → Shareware · “usually free of charge, users can access the source code and alter the program” → Open Source // Free Software. 9618_w24_qp_13_sc_5.a

Tick which licence(s) each statement describes (multi-tick table)

Edit the source code → FSF, OSI · must always pay before use → Commercial · can redistribute → FSF, OSI, Shareware · always gets a trial period → Shareware. 9618_s21_qp_12_sc_2.c

Identify a suitable licence for a scenario and justify the choice

Business needing to modify the source code and paying for maintenance → OSI // FSF: allows the source code to be examined, so the business can adapt it to their needs, and the programmer still earns from the software and updates. 9618_w24_qp_12_sc_5.b.i Game to be sold to the public → Commercial (user pays so the programmer gains income; enables copyright; the user cannot legally edit it; legal action possible against illegal copies) or Shareware (try then pay; more people experience it so more likely to buy). 9618_s23_qp_13_sc_5.b

Benefits of distributing under each licence

Shareware: users can trial and offer feedback · more people try it because the trial is free · bugs found across a wide range of configurations · those who find it useful buy it, giving income · allows the program to be copyrighted. 9618_w23_qp_13_sc_6.a Commercial: the developer earns a fee · comprehensive support is usually offered · the software is usually well tested so has few bugs. 9618_w24_qp_11_sc_7.b Open source: allows users to customise the code · allows errors to be reported and fixed by users · allows additional features to be added · allows collaboration. 9618_w22_qp_13_sc_8.a.i

Explain why a licence might not be appropriate

Open source, for safety-relevant software: the source code would be accessible and could be changed, which might make the program output an incorrect response, so people could purchase the wrong item, with serious consequences (e.g. allergies). 9618_s25_qp_12_sc_3.c.ii

Freeware as a named fifth licence type

Free of charge, but still copyrighted; the developer can set their own restrictions on what a user may do; modification, redistribution or reverse engineering without permission is prohibited. Freeware is credited in 9608 (“name two other types of software licence” → Open Source, Free Software, Shareware, Freeware) but is not named in the 9618 syllabus notes, which list only FSF, OSI, shareware and commercial. Know the distinction — freeware is free but closed; free software is open but not necessarily free of charge — and do not offer freeware where a syllabus-named licence is wanted. 9608_w19_qp_12_sc_5.c.ii

The copyleft condition — redistribution under the same terms

“The source code comes with the software. If the software is modified, the edited source code must be released under the same conditions as the original.” 9608 uses this as the defining feature that separates a Free Software licence from a plain open-source one. The condition appears in no 9618 mark scheme, which says only “users can share the software with others”. 9608_s19_qp_11_sc_2.a, 9608_s21_qp_13_sc_7.b

Benefits to the customer / user rather than the developer

Commercial: potentially better support because a fee is charged · likely fewer bugs / less prone to malware than software distributed under other licences · redress available if the game does not function correctly. Shareware: they can check it works and meets their requirements without paying if it does not. Every 9618 licence-benefit question asks from the programmer’s side. 9608_s20_qp_13_sc_4.c.ii, 9608_s20_qp_13_sc_4.c.iii

Four benefits to a company of choosing commercial software over open source

Support and training readily available · more robust software / fewer bugs, tested more thoroughly and by more users · forums and user groups exist for popular software · an upgrade path is available at minimal cost · the manufacturer develops patches that download automatically · compatibility is inbuilt with other commercial software. 9608_s16_qp_13_sc_3.c

Explain why a developer would not want an open source licence

She does not wish to release the source code · she does not want anyone to edit, modify or share it · she wants to make money from the program. 9618 has asked why open source might not be appropriate for a safety reason (s25_qp_12_sc_3.c.ii); the plain commercial reason is 9608’s. 9608_s19_qp_12_sc_4.d.ii

Drawbacks of each licence

9618 has examined the benefits of shareware, commercial and open source (three separate questions) and only one “why not appropriate”. The drawbacks — open source has no guaranteed support and anyone can alter it; commercial is expensive and locks you in; shareware users may never convert; free software may have no accountable owner — are barely touched in the current series.

A justification question where the obvious answer is wrong

Every 9618 “justify a licence” question has a natural fit (a game → commercial; a modifiable business system → open source). A scenario designed so that the tempting answer is wrong — a charity wanting the widest reach, a safety-critical system, a school with no budget — is the harder version of a question type Cambridge clearly likes.

The five-column comparison

	Commercial	Shareware	Freeware	Open Source (OSI)	Free Software (FSF)
Cost	must be paid for	free trial, then pay for full version	free to use	often free, not always	always free to use
Source code	no	no	no	yes	yes
Can be edited	no	no	no	yes, with conditions	yes, full freedom
Redistribution	no	no	no	yes, under licence terms	yes
Copyrighted	yes	yes	yes	yes	yes
Control	developer retains full	developer retains	developer retains	shared between developers and users	users have full control
The row that matters most: everything is copyrighted, including freeware and open source. Free or open does not mean uncopyrighted — a distinction 9618 tests indirectly through w23_qp_13_sc_6.a (“allows the program to be copyrighted”).

The FSF's four freedoms

1. Freedom to run the program for any purpose · 2. Freedom to study how it works and change it · 3. Freedom to redistribute copies · 4. Freedom to improve it and share the improvements. The FSF promotes freedom, not free price — “free as in libre, not free as in cost” — which is exactly the w25_qp_12_sc_10.b mark point “not necessarily free of charge”, given its underlying logic. Examples: GNU/Linux, LibreOffice, GIMP.

OSI vs FSF, and the crediting condition

The Open Source Initiative promotes software that is freely available and open: users can view, edit and share the source code, encouraging community collaboration and improvement. Open source does not always mean free of charge — some open-source software is paid, but users can still see and change the code. Developers must still credit the original authors when modifying or sharing. Examples: Linux, Firefox, Apache. 9618 treats OSI and FSF as interchangeable in mark schemes (w24_qp_13_sc_5.a accepts either), so know the shared features first and the distinction second.

---

7.1.5 Show understanding of Artificial Intelligence (AI)

Read the syllabus notes exactly

“Understand the impact of AI including social, economic and environmental issues. Understand the applications of AI.” Every 9618 question so far falls into one of two families: how AI works in a described application (usually 3–4 marks) or the impact of AI, by named category (usually 2–4 marks). Note the impact questions always specify which category — “social impacts”, “economic impact”, “ethical impact” — so knowing which bucket an answer belongs in is itself examinable.

Applications of AI — how AI is used in a described system

Image recognition / OCR: explain how AI reads text on an item and translates it

Uses image recognition // optical character recognition · analyses the photograph/pixels to identify the location of characters within the image · converts the pattern of pixels into individual characters · uses natural language to combine characters into words · compares the words to a database to identify a match in the language of the writing · may predict what is missing or incomplete from what has been read · compares to a database of the user’s chosen language to translate · uses text-to-speech to generate the audio waveforms. 9618_s25_qp_12_sc_3.a

Facial recognition: describe how AI identifies a customer / authenticates a face

Captures an image of the face using the camera · uses image recognition · the AI identifies that the image is a face by analysing the pixels to find patterns · measures the distance between facial features · stores the values in a database of user information · is trained to identify the features of a face using a large number of images · compares the calculated distances to those in the database · uses the probability of a match. 9618_s25_qp_11_sc_4.a, 9618_s24_qp_11_sc_5.c.ii

Number-plate recognition: explain how AI identifies a car registration

Uses image recognition · the pixels of each image from the camera are stored and matched to the expected shape/size/colour of a registration plate · uses optical character recognition · the pixels within the identified plate area are analysed and compared to expected characters/letters/numbers. 9618_w24_qp_11_sc_6

Auto-focus on faces: explain how AI focuses a camera on people's faces

Scans the scene in real time · identifies whether there are faces in the image · uses facial/image recognition, taking each frame individually · analyses the pixels · stores the pattern for a face · looks for patterns that match or come close to the pattern for a face · the camera focuses on the pattern identified. 9618_s23_qp_13_sc_3.b

Speech / voice recognition: explain how AI turns speech into commands or orders

Voice/speech recognition is used to identify if someone is speaking · the sound is recorded and analysed · the audio recordings are compared to a database of words/sound waves · to identify the word that has the highest probability of being said · natural language recognition is used · words are combined and compared to known sentences · the programmed actions for matching sentences are performed. 9618_s24_qp_13_sc_7.b, 9618_s23_qp_12_sc_7.d

CCTV / movement tracking: describe how AI identifies and tracks a person

Uses image recognition · monitors every image taken to identify matching images/shapes/features to a “person” · starts recording to secondary storage permanently when a person is identified · identifies the direction of movement and uses it to decide where/how to move the camera · identifies other cameras to start recording based on direction of movement. 9618_s23_qp_11_sc_4.c

Game playing: describe how a computer uses AI to play a board game

The rules / past moves / decision-making algorithms are stored · the AI program is trained by playing many times · it looks ahead at possible moves and/or analyses the pattern of past choices · it chooses the move most likely to be successful · the computer can learn from previous mistakes by storing the positive/negative result of choices and changing its future choices. 9618_s21_qp_12_sc_2.b

The shape of every "how is AI used" answer

Across seven questions the same skeleton earns the marks: capture input → analyse the raw data (pixels / sound waves) → compare against a stored database of patterns → use probability to pick the best match → act on the result → learn and improve from outcomes. Any new application (medical imaging, fraud detection, spam filtering, recommendation systems, autonomous vehicles) is that skeleton with different nouns. Practise fitting it to an unfamiliar scenario rather than memorising the seven above.

Applications not yet used as a scenario

Every examined scenario is image, speech or game. Untested but squarely within “applications of AI”: medical diagnosis, autonomous/self-driving vehicles, fraud detection, recommendation engines, chatbots, spam filtering, weather and climate prediction, robotics in manufacturing. SME’s own impact sections lean on healthcare and climate modelling — both plausible as the setting for a “how does AI do this” question.

Machine learning and training data, as their own question

“The AI program is trained by playing many times” and “trained to identify the features of a face using a large number of images” are credited mark points in s21_qp_12_sc_2.b and s24_qp_11_sc_5.c.ii, but no question has ever asked candidates to explain what training is, why large volumes of data are needed, or what happens when the training data is poor. Given how central it is to every answer above, this is the most likely new AI question.

Why AI rather than a conventional program

No question has asked what makes a system “AI” as opposed to ordinary programming — the ability to learn from data and improve, rather than following fixed instructions. It is the definitional question underlying the whole sub-topic and is untested.

Impact of AI — social, economic, environmental, ethical

Social impacts of students using AI for homework

Negative: using AI to shortcut learning means missing out on developing reasoning and problem-solving skills / not learning anything / decline in specific skills · negative impact on communication, since heavy reliance limits opportunities for collaboration, teamwork and face-to-face communication · emphasises the digital divide, as some students may not have access to the technology · the AI answer might not be correct, leading to mis-information · increased screen time and isolation contributing to anxiety or loneliness. Positive: students given additional support may do better · pupils who struggle in traditional settings may gain self-esteem through AI-assisted learning at their own pace. Long-term: protection of intellectual property / ethical considerations · schools may need to redesign curricula and assessment to integrate AI. 9618_w25_qp_11_sc_8.c

Social benefits of an AI text-reading-and-translating program

It can help people with visual impairment identify which products to purchase · help people who cannot understand the language · help overcome learning or reading difficulties. 9618_s25_qp_12_sc_3.b

Social impact of facial recognition in an airport

Incorrect recognition of faces leads to mistakes, so access to facilities or systems may be denied · privacy issues / people do not like their data being stored · individuals will feel safer · there might be a reduction in crime · faster boarding · catching criminals. 9618_w22_qp_12_sc_9

Economic impact of an AI diagnostic module on a garage

Reduces costs, because less time is taken for diagnosis · increases profits, as technicians spend more time repairing so complete more jobs a day · decreases costs passed to the customer, so the garage may gain customers · profit margins can be reduced, because the program may be expensive to buy, maintain or update. 9618_w22_qp_13_sc_8.b

Ethical impact of AI monitoring students in a classroom

The learning experience could be improved by identifying students who are struggling and employing more personalised learning · there could be early intervention to determine who requires extra support · teachers and students could understand the challenges better and identify optimal times for learning · there could be privacy concerns — students/teachers uncomfortable with constant surveillance, or not wanting their data passed to third parties · there could be mental health concerns due to constant monitoring and pressure, contributing to stress and anxiety. 9618_w25_qp_13_sc_7.a

Environmental impact — the untested third of the syllabus notes

The notes say “social, economic and environmental issues”. Across every 9618 series, questions have asked for social, economic and ethical impacts — never environmental. This is the single clearest gap in Chapter 7 and arguably in the whole AS course: a named category in the syllabus with zero coverage in six years of papers. Prepare it properly (see the !abstract below).

Impacts phrased as a balanced judgement

Every impact question asks for impacts of one polarity or category. “Discuss whether the benefits of AI in this situation outweigh the drawbacks” — requiring both sides and a conclusion — has not been set, though the w25_qp_11_sc_8.c mark scheme (which lists negative, positive and long-term impacts for the same scenario) shows the examiners think in those terms.

Bias in AI decision-making

No 9618 mark scheme mentions bias, yet it is the standard social criticism of AI and follows directly from “trained using a large number of images” — if the training data is skewed, the decisions are skewed. w22_qp_12_sc_9 gets close with “incorrect recognition of faces leads to mistakes”, but the reason is never asked.

Accountability — who is responsible when AI is wrong

Implied by the airport and classroom questions but never asked directly. If an AI system makes a harmful decision, is it the developer, the operator, or the system? This is a genuine ethical question that bridges 7.1.2 and 7.1.5 and is untested.

What AI is, and the two types

AI is a machine that can simulate intelligent behaviours similar to a human: a system that can learn (acquire new information), decide (analyse and make choices) and act autonomously (take actions without human input). Weak AI (narrow AI) — designed to perform a specific task or set of tasks. Strong AI (artificial general intelligence, AGI) — designed to perform any intellectual task a human can do. Neither term appears in the 9618 syllabus notes or any mark scheme, but every application examined so far is weak AI, and the learn/decide/act triple is the cleanest definition for a “what is AI” question that has never been set.

The three characteristics of an AI system

Collection of data — AI systems require large amounts of data, processed using rules or algorithms that let the system make decisions and predictions · Rules for using data — logical reasoning to evaluate information and make decisions based on it · Ability to reason — it can change its own rules and data. The third is the important one: self-modification is what separates AI from a conventional program, and it is the theory behind “the computer could learn from previous mistakes … and change its future choices” in s21_qp_12_sc_2.b.

Headline advantages and disadvantages

Advantages: increased efficiency · increased accuracy · scalability. Disadvantages: job losses · potential for biased decision making · ethical concerns over its use. The bias point is not in any mark scheme and fills the !info gap above.

Social impact — the full picture

Workforce: automation may replace some roles, causing unemployment or changed job roles; at the same time new jobs emerge requiring AI knowledge or human–AI collaboration; reskilling and upskilling programmes are essential. Education and accessibility: those with better technology, education and internet access benefit more, creating a digital divide; equal access to AI education, tools and training matters. Healthcare: AI can improve diagnosis, treatment planning and patient monitoring, but is not infallible — a wrong diagnosis can have serious consequences; the extent of human oversight must be determined, since decisions made solely by AI in critical situations are risky; and who is responsible when something goes wrong (developer, system or healthcare provider) creates legal and ethical challenges needing clear guidelines and regulation. The healthcare accountability paragraph is the answer to the !info accountability gap, and the digital-divide point is one 9618 does credit (w25_qp_11_sc_8.c).

Economic impact — the full picture

Employment and industry: job displacement in sectors relying on routine or manual work · increased productivity and efficiency in manufacturing, logistics and finance · growing demand for AI-related roles such as data scientists and machine learning engineers · governments and businesses must invest in retraining. Business and innovation: enables new business models (personalised services, automated customer support) · supports faster innovation by improving R&D · reduces operational costs through automation and predictive analytics · but small businesses may struggle to compete with large companies that can invest more in AI, widening inequality between organisations. Market dynamics: AI can concentrate wealth and power in large tech companies that control the key tools and data, creating monopolistic advantages and reducing competition, so policymakers must consider regulation to ensure fair access and prevent deepening income inequality. The only economic question in the series (w22_qp_13_sc_8.b) stays at the level of one garage’s costs and profits. Everything above is the macro version, untested.

Environmental impact — the content for the biggest gap in the chapter

Energy consumption: AI systems, especially large-scale models, require vast computing power, leading to high electricity usage and significant carbon emissions, and can strain power grids if deployed at scale without renewable support. Mitigations: optimise models to be more energy-efficient; use green data centres powered by renewables. Climate modelling and sustainability: AI is also a tool for environmental protection — climate modelling, predicting weather patterns, analysing environmental data, optimising energy use in smart grids and buildings, and supporting sustainable agriculture by analysing soil, weather and crop data to reduce waste. The positive applications must be weighed against AI’s own resource demands. E-waste and hardware lifespan: AI drives demand for specialised hardware (GPUs, TPUs), shortening device lifespans through rapid advancement and increasing electronic waste, which pressures recycling systems. Sustainable hardware design, recycling and component reuse are needed. Environmental impact is named in the syllabus notes and has never been examined in any series. A two-sided structure — AI’s own footprint (energy, emissions, e-waste) versus AI as an environmental tool (climate modelling, energy optimisation, agriculture) — is the safest preparation, and mirrors how the examined social and economic questions are marked.