Social, Economic, and AI Impacts

You can read this note directly if you understand that technology changes how people communicate, work, learn, shop, and make decisions.

This note focuses on impact reasoning. In H2 Computing, this usually means explaining how a computing system affects different people, organisations, and society.

A strong answer should usually include:

• a named issue;
• an affected stakeholder;
• a benefit;
• a risk or harm;
• a suitable safeguard;
• a balanced judgement.

Beginner Problem

Impact questions are rarely one-sided.

Example:

An online learning platform lets students watch lessons at home.

Benefit:

• students can revise at their own pace.

Risk:

• students without stable internet or a quiet study space may be disadvantaged.

A good answer explains both sides and then suggests a safeguard, such as providing offline materials or school-based access.

Weak answer:

Online learning is good because students can study anywhere.

Stronger answer:

Online learning can make revision more flexible because students can replay lessons
at their own pace. However, students without reliable internet, a suitable device,
or a quiet study space may be disadvantaged. The school could reduce this risk by
providing offline materials, loan devices, or school-based access.

The stronger answer is better because it names a benefit, a disadvantaged group, and a practical safeguard.

Social Impact

Social impact means the effect on people, communities, behaviour, relationships, access, trust, and well-being.

Examples:

Technology	Possible benefit	Possible risk
social media	fast communication and community building	cyberbullying, misinformation, anxiety
online learning	flexible access to lessons	unequal access to devices or quiet study space
health apps	easier self-monitoring	sensitive data may be exposed
facial recognition	faster identity checks	surveillance, bias, loss of privacy
public-service chatbot	quicker responses to common questions	users may receive misleading responses or be unable to challenge decisions
AI proctoring	supports remote assessment	privacy concerns, false suspicion, stress

Do not write only benefits or only harms. Strong answers discuss trade-offs.

Useful social-impact questions:

• Does the system change how people communicate or behave?
• Does it affect privacy, trust, safety, or well-being?
• Does it help some groups more than others?
• Are vulnerable or less digitally confident users affected?
• Can affected users understand, challenge, or opt out of the system?

Digital Divide

The digital divide is unequal access to technology, connectivity, or digital skills.

It can involve:

• no suitable device;
• poor internet connection;
• lack of assistive technology;
• low digital literacy;
• language or accessibility barriers;
• lack of confidence using digital forms or online services.

Example:

Moving a government service fully online may improve efficiency, but it may disadvantage people who lack devices, internet access, or confidence using digital forms.

Possible safeguards include:

• keeping an offline or assisted service channel;
• providing help desks or guided support;
• making the interface accessible;
• using clear language;
• allowing trusted caregivers or staff to assist where appropriate.

Exam tip:

Do not only say "some people cannot use technology". Name the barrier:
no device, weak internet, low digital literacy, disability access, language barrier,
or lack of confidence.

Economic Impact

Economic impact means effects on jobs, income, productivity, costs, markets, and business models.

Examples:

Impact	Explanation
productivity gain	software automates repetitive tasks
job displacement	some routine roles need fewer workers
new jobs	demand grows for data, cybersecurity, AI, and software skills
reskilling need	workers need training for changed roles
lower transaction cost	e-commerce and digital payment reduce friction
inequality risk	gains may benefit some groups more than others
platform economy	digital platforms create new ways to sell, deliver, or freelance
market concentration	large platforms may gain strong control over users, data, and sellers

Automation is not simply “good” or “bad”. It depends on who benefits, who bears the cost, and what support is provided.

Example:

Self-checkout machines may reduce labour cost and shorten queues for some customers,
but they may reduce demand for cashier roles and make shopping harder for elderly
or disabled customers. A responsible deployment should retain staff assistance,
train workers for new roles, and monitor whether customers are excluded.

Stakeholder Analysis

Caption: Stakeholder analysis helps avoid one-sided answers.

Example: self-checkout machines in a supermarket.

Stakeholder	Benefit	Risk
customers	shorter queues for some purchases	difficulty for elderly or disabled users
workers	fewer repetitive checkout tasks	reduced cashier roles
business	lower operating cost	customer frustration if system fails
society	productivity improvement	employment transition pressure

When writing about stakeholders, avoid saying only “people are affected.” Name the group and the effect.

Weak:

Automation affects people.

Stronger:

Cashiers may need retraining because self-checkout machines reduce the number
of routine checkout tasks, while customers may benefit from shorter queues.

Useful stakeholder groups include:

Scenario	Possible stakeholders
school app	students, parents, teachers, school, vendor
health app	patients, doctors, clinic, insurer, app provider
AI hiring system	applicants, HR staff, company, rejected candidates
social media platform	users, moderators, advertisers, public, vulnerable groups
workplace monitoring	employees, managers, company, customers

AI Ethics

AI systems can support prediction, classification, recommendation, generation, and automation. Their risks must be managed.

Key AI ethics issues:

Issue	Meaning
bias	system performs unfairly for some groups
transparency	users may not know how a decision was made
accountability	unclear who is responsible for harm
privacy	training, input, or usage data may include sensitive information
safety	incorrect output may cause real harm
human oversight	important decisions may need human review
over-reliance	users may trust AI output without checking it
contestability	affected users may need a way to question or appeal a decision

Example:

An AI tool shortlists job applicants.

Benefits:

• faster screening;
• consistent processing of large applicant pools.

Risks:

• biased training data may disadvantage groups;
• applicants may not know why they were rejected;
• over-reliance may remove meaningful human judgement.

Safeguards:

• audit outcomes for bias;
• explain decision factors where practical;
• keep human review for high-stakes decisions;
• protect applicant data;
• monitor system performance after deployment;
• allow applicants to request review where appropriate.

Beginner checkpoint: AI risk is not only about whether the program is accurate. A system can be accurate on average but still unfair to a particular group.

AI Impact Examples

AI proctoring in school

Possible benefits:

• supports remote assessment;
• may reduce the need for physical invigilation;
• can flag suspicious behaviour for review.

Possible risks:

• students may feel constantly watched;
• false positives may unfairly suspect honest students;
• camera, microphone, or screen data may create privacy concerns;
• students with poor internet, shared rooms, disabilities, or anxiety may be disadvantaged.

Safeguards:

• tell students clearly what is recorded and why;
• collect only necessary data;
• keep human review before penalties;
• provide alternatives for students with access or disability needs;
• delete recordings when no longer needed.

AI recommendation system

Possible benefits:

• users find relevant content more quickly;
• businesses can personalise services;
• platforms can increase engagement.

Possible risks:

• users may be trapped in filter bubbles;
• misinformation may spread if engagement is rewarded over accuracy;
• vulnerable users may be shown harmful content;
• users may not know why certain content is promoted.

Safeguards:

• give users control over recommendations;
• down-rank harmful or false content;
• label sponsored or AI-recommended content clearly;
• allow users to report harmful content;
• audit the system for unintended harm.

AI chatbot for public services

Possible benefits:

• citizens can get faster answers;
• staff can focus on complex cases;
• service can be available outside office hours.

Possible risks:

• wrong answers may mislead users;
• users may not know whether advice is official or approximate;
• elderly or less digitally confident users may struggle;
• personal data typed into the chatbot may be mishandled.

Safeguards:

• make clear when users are interacting with AI;
• provide a route to human assistance;
• avoid asking for unnecessary personal data;
• log and review common errors;
• state that high-impact decisions require official confirmation.

Misinformation and Online Harm

Computing systems can spread information very quickly. This creates social, ethical, and sometimes legal concerns when false, abusive, or harmful content reaches many people.

Common issues:

Issue	Meaning	Possible harm
misinformation	false or misleading information is shared	public confusion, reputational harm, unsafe decisions
harassment	repeated abusive, threatening, or insulting behaviour	emotional distress, fear, unsafe online environment
cyberbullying	online bullying, often among peers or students	stress, anxiety, social isolation
doxxing	publishing private information to expose or intimidate someone	privacy loss, harassment, physical safety risk
deepfake or manipulated media	generated or edited media presents false impression	deception, reputation damage, loss of trust

At H2 level, the important skill is not to memorise legal sections. The important skill is to recognise the issue, identify the harm and stakeholder, and propose responsible action. Online falsehoods, harassment, doxxing, personal data, and platform policy may be relevant depending on the scenario.

Example:

A student creates a fake image of a teacher and shares it online as if it were real.

Possible points:

• the teacher may suffer reputational and emotional harm;
• other students may believe false information;
• the school community may lose trust;
• the platform or school may need to remove the content and investigate;
• responsible action includes reporting the post, preserving evidence, removing harmful content, and educating students about online harm.

Exam tip:

For misinformation or harassment scenarios, do not only say "it is wrong".
Explain the harm, affected stakeholder, and suitable response.

Issue Classification

Caption: A single scenario can raise several issue types.

Example: an AI surveillance camera in a workplace.

Issue type	Possible point
ethical	employees may feel constantly monitored
legal	personal data and workplace monitoring rules may apply
social	workplace trust may decrease
economic	productivity may improve but morale may fall

Worked classification:

Scenario: A company uses software to monitor employee keyboard activity.

Category	Possible point
ethical	monitoring may be excessive if employees are not properly informed
legal	workplace data collection and personal data rules may apply
social	trust between staff and management may decrease
economic	productivity may improve, but staff morale and retention may worsen

This shows why one scenario can produce several valid discussion points.

Another worked classification:

Scenario: A social media platform uses AI to recommend news videos to teenagers.

Category	Possible point
ethical	the platform should avoid promoting harmful or misleading content simply to increase engagement
legal	online falsehoods, personal data, or harassment concerns may be relevant depending on the case
social	teenagers may form distorted views if they repeatedly see one-sided or false content
economic	the platform earns advertising revenue from engagement, which may conflict with user well-being

Exam Answer Toolkit

Different command words require different answer depth.

Command word	What to do	Example response style
identify	name the issue or stakeholder	”One social issue is the digital divide.”
describe	state what the issue means in the scenario	”Some users may lack reliable internet access.”
explain	give cause and consequence	”This disadvantages them because they cannot access the service when it becomes fully online.”
discuss	give both benefit and risk	”The system improves efficiency, but may exclude less digitally confident users.”
evaluate	reach a balanced judgement with conditions	”It is acceptable only if offline support and data protection are provided.”

A simple structure for longer answers:

Step	What to include
1. Context	What technology is being used?
2. Stakeholder	Who is affected?
3. Benefit	What improvement does the technology bring?
4. Risk	What harm or unfairness may occur?
5. Safeguard	What can reduce the risk?
6. Judgement	Under what conditions is it acceptable?

Useful phrases:

This may benefit ... because ...
However, it may disadvantage ... because ...
A suitable safeguard is ...
The organisation should ...
This is acceptable only if ...
Human review is needed when ...
Users should be informed about ...

Worked Exam-Style Answers

2-mark identification answer

Question:

A school moves all administrative forms online. Identify one possible social issue.

Answer:

One social issue is the digital divide. Some parents or students may lack reliable
internet access, suitable devices, or confidence using online forms.

Why this works:

• names the issue;
• links it to the scenario.

4-mark discussion answer

Question:

A supermarket introduces self-checkout machines. Discuss one economic benefit
and one economic risk.

Answer:

One economic benefit is improved productivity, because fewer staff may be needed
for routine checkout tasks and queues may move faster. However, an economic risk
is job displacement, because cashier roles may be reduced or changed. The business
should provide retraining so affected workers can move into customer-support,
inventory, or service roles.

Why this works:

• explains benefit;
• explains risk;
• adds a safeguard.

6-mark evaluation answer

Question:

A company uses an AI system to shortlist job applicants. Evaluate whether this is
responsible.

Answer:

The system may be useful because it can process many applications quickly and apply
consistent screening criteria. However, it may be unfair if the training data reflects
past biased hiring patterns, causing some groups of applicants to be disadvantaged.
Applicants may also not know why they were rejected, which creates a transparency
and accountability issue.
 
The company should audit the system for biased outcomes, protect applicant data,
explain the main screening criteria where practical, and keep human review for
borderline or high-impact decisions. Therefore, using AI for shortlisting may be
responsible only if it supports human judgement rather than replacing accountability.

Why this works:

• gives benefit;
• gives multiple risks;
• gives safeguards;
• reaches a balanced judgement.

Common Mistakes

• Assuming all automation causes unemployment in the same way.
• Ignoring reskilling and new job creation.
• Saying AI is neutral because it is a machine.
• Discussing privacy but forgetting fairness or accountability.
• Treating “more efficient” as automatically better for everyone.
• Forgetting vulnerable or less digitally confident users.
• Saying “people are affected” without naming the stakeholder group.
• Giving a safeguard that is too vague, such as “make it secure”, without saying how.
• Naming a law without linking it to the facts of the scenario.
• Writing only personal opinion instead of explaining benefit, risk, and condition.

Check Your Understanding

1. What is the digital divide?
2. Give one economic benefit and one economic risk of automation.
3. Why can AI systems be biased?
4. Why is stakeholder analysis useful?
5. Why can misinformation be a social harm?
6. Give one safeguard for AI use in high-impact decisions.

Answers:

1. Unequal access to digital technology, connectivity, or skills.
2. Benefit: improved productivity. Risk: displacement of some routine jobs.
3. Training data, design choices, or deployment context may favour some groups over others.
4. It helps identify benefits and harms for different affected groups.
5. It may mislead people, damage reputation, reduce trust, or cause unsafe decisions.
6. Keep human review, audit for bias, explain decision factors where practical, and allow review or appeal where appropriate.

Related Notes

• Social, Ethical, Legal, and Economic Issues
• Professional Ethics and Legal Responsibilities