Big Idea 5: Impact of Computing

Computing Innovations: Effects, Trade-offs, and Unintended Consequences

A computing innovation is a new or improved computer-based product or system (often including hardware, software, and data working together) that changes how people live, work, or communicate. “Innovation” does not always mean a brand-new invention; it can be a powerful recombination of existing technologies. For example, smartphones combine GPS, cameras, mobile internet, and apps to create new capabilities.

When you’re asked about the impact of computing, you’re being asked to reason about how an innovation changes the world. Strong explanations usually do three things: they identify who is affected (individuals, communities, organizations, governments, the environment), explain what changes (behavior, access, efficiency, power, opportunity, risk), and acknowledge trade-offs, because most innovations create benefits and harms at the same time.

New ways to communicate and interact

Thanks to the Internet and the ease of collaboration and sharing, programs and apps (software applications) can be quickly and easily shared with people worldwide. That fast spread can create huge impacts (positive or negative), and it can also produce results not originally foreseen by the software developers. In practice, programmers and businesses try to identify potential negative uses, but it is seldom possible to think of all the ways other people could use an innovation.

Several widely used examples of communication-changing innovations include:

• The World Wide Web, which was originally designed for scientists to share their research.
• Social media, which has been used to stream events across the globe and has sometimes helped to change history.
• Online learning, an education model that would not be possible without the tools of communication available via the Internet.
• Machine learning and data mining, which help find patterns and identify insights in data, leading to new innovations.

Beneficial vs. harmful effects (and why both can be true)

A beneficial effect is an outcome that improves quality of life, access, safety, productivity, or knowledge. A harmful effect is an outcome that creates risk, inequity, damage, exploitation, or loss (including privacy loss). A key skill in AP CSP is avoiding one-sided thinking: the same innovation can help one group and harm another, or help in one way while harming in another.

For example, consider a navigation app:

• Beneficial: reduces travel time, helps emergency response, increases accessibility for new drivers.
• Harmful: encourages constant location tracking, can reroute traffic into residential neighborhoods, can be misused for stalking.

Direct vs. indirect impacts

It helps to separate effects into:

• Direct effects: immediate, intended results (for example, video conferencing enables remote meetings).
• Indirect effects: secondary consequences that emerge later (for example, remote work changes housing demand, commuting patterns, and local business revenue).

AP-style prompts often reward you for going beyond the obvious direct effect and explaining at least one plausible indirect effect.

How innovations spread and reshape society

Computing innovations scale quickly because digital information is easy to copy and distribute. Once a system becomes widely adopted, network effects can occur, meaning the innovation becomes more valuable as more people use it (for example, messaging platforms). This can lead to rapid adoption, market concentration (a few companies dominate), and new norms (expectations about availability, communication speed, or “being online”).

Worked example: Social media as a computing innovation

To evaluate impact, first describe what the system includes:

• Software platforms (apps, feeds)
• Algorithms (ranking, recommendations)
• Data collection (likes, clicks, time watched)
• Infrastructure (servers, networks)

Then evaluate impacts:

• Beneficial: community building, disaster communication, small-business marketing, political organizing.
• Harmful: misinformation spread, harassment, addictive design patterns, mental health concerns, privacy loss through tracking.

A key insight is that many “social impacts” are driven by design choices in software and data use (especially algorithmic ranking and data collection).

Exam Focus

• Typical question patterns:
  • Describe one beneficial and one harmful effect of a given computing innovation (often for different stakeholders).
  • Explain how a computing innovation changes communication, work, education, health, or civic life.
  • Identify a plausible unintended consequence of widespread adoption.
• Common mistakes:
  • Giving only a benefit (or only a harm) when the question asks for both.
  • Stating impacts vaguely (“it helps people”) without explaining who, how, and why.
  • Confusing the innovation with a single device; many innovations are socio-technical systems (technology plus people plus policies).

Access to Information: The Web, Cloud Computing, and Open Data

Modern computing dramatically increases access to information, and that access reshapes learning, markets, culture, and problem-solving. The same underlying theme appears again and again: information becomes easier to store, search, share, and remix—often at global scale.

Cloud computing

Cloud computing offers new ways for people to communicate, making collaboration easier and more efficient. Storing documents in the “cloud” means they are stored on a computer server at a location different than where the owner of the files is located.

Open access data and public databases

The sharing of huge amounts of public data by organizations (such as the U.S. government) provides the opportunity for anyone to search for information or help solve problems. The availability of open databases in many fields—including science, entertainment, sports, and business—has benefited people everywhere.

Search trends and analytics

Social media sites and search engines often publish what the most frequent searches and posts are about. Browsers may keep a list of your most frequently visited sites on their home page to help you out. Search engines can also identify when more people than usual are watching a video or searching for a topic.

Analytics identify trends for marketing purposes and help businesses determine what and where customers are searching for (their products and their competitors’), how long an item sits in a virtual shopping cart, and when people buy.

Targeted advertising

Targeted advertising can be helpful for businesses and consumers when looking for a specific item. At the same time, targeted advertising often depends on collecting and analyzing user data, which connects directly to privacy risks and power imbalances discussed in the next section.

Exam Focus

• Typical question patterns:
  • Explain how cloud computing changes collaboration or access to files.
  • Describe a beneficial impact of open data (research, accountability, innovation) and a possible risk (misuse, privacy concerns if data can be linked).
  • Connect search trends/analytics or targeted advertising to data collection and user impact.
• Common mistakes:
  • Treating “more information” as automatically good without considering misuse, privacy, or unequal access.
  • Describing cloud storage as “on your computer” rather than on remote servers.

Data and Privacy: What Gets Collected, What Can Be Inferred, and Why It Matters

Modern computing runs on data. Data are facts or measurements collected for reference or analysis. For impact questions, the key idea is not just that data exists, but that data collection changes power relationships: whoever collects and controls data can influence decisions, markets, and people.

Personally identifiable information (PII) and inference

Personally identifiable information (PII) is any information that identifies you directly or indirectly. It includes obvious identifiers like your name and address, but also data such as your age or social security number, and it can include sensitive categories like medical or financial information.

A major privacy concept is inference: even if a dataset does not include a person’s name, it may still be possible to figure out who they are or learn sensitive facts about them.

• “Anonymous” location points can reveal home and work locations.
• Purchase history can suggest health conditions or religious/political affiliation.

Websites may also use PII and related browsing data to show you certain information or related topics based on your prior visits.

How data is collected in real systems

Data collection can be:

• Explicit: you type it in (sign-up forms, surveys).
• Implicit: the system records behavior (clicks, watch time, location, device identifiers).

Many people willingly provide personal information to sites to gain access or privileges (sports teams, shopping, restaurants). Their data is stored and may be sold with or without their knowledge or permission.

Digital footprints and fingerprints

A digital footprint (sometimes described along with digital fingerprints) is the trail of little pieces of data you leave behind as you go through your daily life online—posts, messages, photos, logins, and metadata. The impact is that digital information is easy to copy and redistribute and difficult to fully erase.

Some web browsers offer “incognito” or “private” modes so that web searches and file downloads are not recorded in the web history on that device. Some browsers also attest that they do not track and retain your search data.

Data use: personalization, prediction, and decision-making

Once collected, data can be used to:

• Personalize content (recommendations)
• Predict behavior (likelihood you’ll click, buy, or churn)
• Automate decisions (loan approvals, hiring filters)

These uses can improve convenience and efficiency, but they can also produce unfairness, manipulation, or privacy harm—especially when people do not understand what is being collected or how it is used.

Anonymization and why it can fail

Anonymization removes personal identifiers from data, but it is not a guarantee of privacy. Datasets can often be re-identified by linking them with other datasets (for example, matching a few location/time points to a known person). A strong takeaway is that anonymization reduces risk but does not eliminate it.

Surveillance and tracking

Surveillance is monitoring behavior or activities, often at scale. Computing enables surveillance through cameras and face recognition, location tracking via phones, and online tracking via cookies and device fingerprints. Surveillance can support fraud prevention, missing-person searches, and public safety, but it can also enable authoritarian abuse, create chilling effects on speech, support discrimination, or facilitate stalking.

Data breaches and secondary harms

A data breach occurs when private data is accessed or exposed without authorization. The impact goes beyond embarrassment: breaches can cause identity theft, financial fraud, targeted scams (phishing), and long-term loss of privacy because copied data can persist indefinitely.

Concrete example: Fitness trackers

A fitness tracker may collect heart rate, sleep patterns, location, and exercise routines.

• Beneficial: helps you build healthy habits; can support medical monitoring.
• Harmful: could expose sensitive health inferences; location trails can reveal daily routines; insurers or employers might pressure people to share data.

This illustrates a common theme: data collected for one purpose can later be used for another, sometimes without meaningful consent.

Exam Focus

• Typical question patterns:
  • Explain a privacy risk from a computing innovation that collects data.
  • Describe how data can be used to infer sensitive information.
  • Identify a trade-off between personalization and privacy.
• Common mistakes:
  • Treating “anonymous” data as automatically safe without discussing re-identification risks.
  • Saying “just don’t share data” without recognizing implicit collection and unavoidable digital traces.
  • Confusing privacy (control over personal info) with security (protection against unauthorized access). They are related but not identical.

Digital Divide and Accessibility: Who Gets Included, Who Gets Left Out, and How to Design for Everyone

The digital divide refers to unequal access to computing devices, the internet, and the knowledge needed to use them effectively. Technology has had a major impact on the world by enabling innovation through the sharing of resources and computational artifacts, allowing people to virtually meet from anywhere, and helping push society toward being more globally connected. However, unequal access changes who benefits from these possibilities.

Dimensions of the digital divide

It helps to think of at least three layers:

1. Access divide: Who has reliable devices and internet?
2. Use divide: Who has the skills, time, and support to use technology effectively?
3. Outcome divide: Who benefits from technology (jobs, education, health), and who does not?

The impact of the digital divide includes unequal access to information, knowledge, markets, and different cultures. For example, someone might have a phone but no laptop, or internet that is too slow for video classes—both still count as part of the divide.

Why the digital divide matters

Computing is a gateway to opportunity. When access is unequal, computing can amplify existing inequality in education (remote learning, research tools, tutoring platforms), employment (online applications, remote work, training), and civic life (access to government services and information).

Factors that contribute to unequal access

Common contributors include income and cost of devices/internet, geography (rural broadband availability), language and cultural barriers, disability and lack of assistive technology, and policy decisions and infrastructure investment.

Accessibility and assistive technology

Accessibility means designing computing innovations so that people with disabilities can perceive, understand, navigate, and interact with them. Assistive technology (hardware or software that helps people perform tasks) can be specialized (Braille displays) or mainstream (voice typing), and it often benefits many users beyond the originally targeted group.

Examples of accessibility features include screen readers, captions, keyboard navigation and switch controls, high-contrast modes, and scalable text. A key mindset is that disability is often a mismatch between a person and a design environment; better design reduces that mismatch.

Inclusive design

Inclusive design means planning for diverse users from the start rather than “patching” accessibility later. Retrofitting accessibility can be expensive and incomplete, so early planning matters for equity.

Example: Online testing platform

If a school adopts an online testing system:

• Benefit: faster grading, flexible scheduling.
• Potential harm: students without reliable internet or devices are disadvantaged; students needing screen readers may face compatibility issues; time limits may not account for accommodations.

This highlights that impact depends on context: the same tool can be fair in one environment and unfair in another.

Exam Focus

• Typical question patterns:
  • Describe how unequal access to computing affects education, jobs, or civic participation.
  • Explain how an innovation could reduce or worsen the digital divide.
  • Identify an accessibility feature that addresses a given barrier.
• Common mistakes:
  • Treating the digital divide as only “having internet or not,” ignoring quality, skills, and outcomes.
  • Assuming technology automatically increases equity without considering cost, language, disability, and infrastructure.
  • Confusing accessibility (designing for disability) with general usability (designing for convenience). They overlap but are not the same.

Bias in Computing: How Data and Algorithms Can Reinforce Inequality

Bias is a systematic tendency toward certain outcomes. It can be described as intentional or unintentional prejudice for or against certain groups of people, and it shows up in computing innovations too. Bias is not always intentional; a system can produce unfair outcomes even if no one “meant” to discriminate.

Where bias comes from

Bias can enter through biased data (unrepresentative datasets), biased labels or measurements (human judgments encoded into what’s measured), biased model goals (optimizing for profit or accuracy without considering fairness), and biased deployment context (high-stakes settings like policing, hiring, and credit).

Because humans write algorithms, our biases can make their way into the algorithms and the data used by innovations without us realizing it.

Representation and sampling problems

A common issue is underrepresentation, where some groups appear less in the dataset, so the system learns patterns that work better for the majority.

Example: a facial recognition system trained mostly on lighter-skinned faces may perform worse on darker-skinned faces.

Feedback loops

A feedback loop occurs when a system’s outputs influence future inputs, reinforcing patterns over time.

Examples:

• Recommendation systems promote certain content; higher views teach the algorithm to recommend that content even more.
• Predictive policing sends more patrols to a neighborhood; more recorded incidents then “justify” even more patrols.

High-stakes uses of AI

Artificial intelligence programs are increasingly used for screening job candidates, determining whether a person merits credit to purchase a house, and locating what areas have more crime. These uses can scale bias, especially when decisions are automated without transparency and accountability.

Why “the computer decided” is not an excuse

Algorithmic outputs reflect human choices (what to optimize), historical data (which can encode past discrimination), and constraints (imperfect measures). Algorithms can appear neutral while still producing unequal outcomes.

Fairness, transparency, and accountability

Mitigations include improving data representation, testing outcomes across subgroups (not just overall accuracy), providing explanations where possible, and adding human oversight for high-stakes decisions.

Example: Automated resume filtering

A company uses software to rank applicants.

• Potential benefit: faster processing of many applications.
• Potential harm: if past hires were mostly from certain schools or demographics, the training data may reflect that history; the system may penalize nontraditional experiences; qualified candidates could be filtered out.

Exam Focus

• Typical question patterns:
  • Explain how a dataset can cause biased results in an algorithmic decision system.
  • Describe a harmful effect of using an algorithm in hiring, lending, or policing.
  • Suggest a way to reduce bias (better data, auditing, transparency, oversight).
• Common mistakes:
  • Claiming “algorithms are unbiased because computers are objective.”
  • Talking only about “bad programmers” instead of structural issues like data representation and feedback loops.
  • Offering unrealistic fixes (“remove all bias”) rather than describing practical mitigation steps.

Crowdsourcing and Collective Computing: Power, Quality, and Ethics

Crowdsourcing is getting contributions (data, ideas, labor, money, or computing power) from a large group of people, typically via the internet. It changes how work is organized and how knowledge is created.

Why crowdsourcing works

Crowdsourcing can succeed because many people contribute small pieces that add up, diverse participants bring different perspectives, and tasks can be parallelized.

Types and examples of crowdsourcing

Crowdsourcing appears in many forms:

• Crowdfunding: raising money from many people.
• Citizen science: volunteers contribute observations or analysis.
• Human computation / microtasking: people do small tasks computers struggle with (labeling images, transcribing audio).
• Collaborative knowledge building: many people create and edit shared resources.

Crowdsourcing can also mean asking the “crowd” (anyone who accesses a site) for feedback to solve problems, find employment, or secure funding.

A related form is contributing computing resources: scientists may share data and ask “citizen scientists” to look for patterns or to donate computer time while a machine is inactive. This can “scale up” processing capability at little to no cost to the organization seeking resources.

Quality control and misinformation risks

Crowdsourcing can fail due to unrepresentative contributors (sampling bias), incentives that encourage low-quality work, or malicious manipulation. Common safeguards include reputation systems, redundancy (multiple contributors do the same task), and moderation/review.

Labor and ethics

Crowdsourcing can create flexible opportunities and global participation, but it can also involve low pay and lack of protections, unclear consent about how contributions will be used, and unequal distribution of benefits between platforms and contributors.

Example: Crisis mapping during disasters

During a natural disaster, volunteers may update maps with road closures, shelter locations, and damage reports.

• Beneficial: rapid information updates; helps responders allocate resources.
• Harmful: incorrect reports can mislead; sharing exact locations might endanger vulnerable people; uneven participation can leave some areas unmapped.

Exam Focus

• Typical question patterns:
  • Explain how crowdsourcing enables a project that would be difficult otherwise.
  • Identify a risk of crowdsourcing (quality, bias, manipulation, labor ethics).
  • Compare benefits and drawbacks for different stakeholders.
• Common mistakes:
  • Describing crowdsourcing only as “people helping” without addressing verification and incentives.
  • Ignoring ethical concerns about labor, consent, or exploitation.
  • Assuming more contributors automatically means better outcomes.

Legal and Ethical Concerns: Intellectual Property, Licensing, and Responsible Use

Computing makes copying and distributing information extremely easy, creating tension between openness and the rights of creators and users.

Intellectual property and copyright

Intellectual property refers to creations of the mind (music, writing, software, images). A key principle is that anything a person creates, including computational artifacts, is that person’s intellectual property. Copyright is a legal protection that gives creators rights over how their work is used and distributed.

Because digital data is easy to duplicate perfectly and instantly, copyright issues show up frequently. “It’s online” does not mean “it’s free to use however you want.” Material created by someone else that you use in any way should always be cited.

Peer-to-peer networks and illegal sharing

Peer-to-peer networks exist that are used to illegally share files of all types. This is a common example of how frictionless copying can create legal and ethical conflicts.

Licensing and Creative Commons

A license is a legal agreement describing how something can be used (installed, modified, redistributed, remixed). Creative Commons provides a way for creators of software, images, music, videos, and any computational artifact to share their creations with stipulations for sharing clearly indicated (such as attribution or noncommercial use). Because digital data is easy to find, copy, and paste, ensuring you have written permission from the creator or owner (or a license that grants permission) is important.

Open-source software

Open-source software makes source code available under a license that allows people to view, modify, and share it under specific conditions. It is often freely shared, updated, and supported by anyone who wants to do so. Open source can accelerate innovation, improve transparency, and expand people’s abilities to participate in tasks they might not otherwise be able to do—but it still comes with rules defined by the license.

Ethical computing beyond legality

Ethics concerns what is right and fair. Something can be legal but unethical (for example, designing an app to maximize addictive engagement in children). Ethical reasoning asks who might be harmed even if the system is profitable or popular, and it often involves harm prevention, honesty, respect for privacy and consent, and fairness and inclusion.

Data ownership, consent, and always-on monitoring

A recurring issue is who “owns” data about you and whether people meaningfully consented to collection and use. Devices that continually monitor and collect data—such as voice-activated devices installed in homes or facial-recognition cameras posted in communities—can raise legal and/or ethical issues.

Example: Using an image in a school app

If you want to use an image you found online in an app interface, responsible questions include: Is it copyrighted? Is it under a license that allows reuse? Do you need attribution? Would a public-domain or properly licensed image be a better choice?

Exam Focus

• Typical question patterns:
  • Distinguish between legal and ethical considerations in a scenario.
  • Explain why copying digital content raises intellectual property concerns.
  • Identify appropriate actions (attribution, permission, using open licenses).
• Common mistakes:
  • Assuming educational use always removes copyright concerns.
  • Treating “open source” as “no rules.”
  • Reducing ethics to “don’t hack,” instead of considering manipulation, bias, and privacy.

Safety, Security, and Cybersecurity: Protecting People in a Connected World

Big Idea 5 emphasizes how unsafe systems and unsafe behavior harm individuals and communities, and what responsible participation looks like.

Security vs. privacy (connected but different)

• Security is protecting systems and data from unauthorized access or damage.
• Privacy is controlling personal information and how it is used.

A system can be secure but privacy-invasive (securely storing detailed location histories), or it can collect minimal data but be insecure (weak passwords).

Protecting our data and the global impact of cybersecurity

Many aspects of life are easier because the Internet provides easy access to shopping, entertainment, sports sites, and price comparisons. At the same time, cybersecurity has a global impact because anyone from anywhere can attempt to gain unauthorized entry to someone else’s computer, data, servers, or network.

Security practices: strong passwords and multifactor authentication

The security of your data includes preventing unauthorized individuals from gaining access and preventing those who can view data from changing it. Strong passwords help block unauthorized access, and multifactor authentication adds an increasingly common extra layer of protection.

Common threats: phishing, malware, viruses, keylogging, and social engineering

Cybersecurity protects devices and networks from attacks and unauthorized use. Different attacks cause different problems: data may be damaged, or a device may be used to further spread malware.

Common threats include:

• Phishing: emails and/or websites that look legitimate in hopes of inducing someone to click a malicious link.
• Malware: harmful software that can steal data, damage files, or take control of devices.
• Computer viruses: like human viruses, they attach themselves to (or are part of) an infected file.
• Keylogging software: a form of malware that captures every keystroke and transmits it to whoever planted it.
• Social engineering: manipulating people rather than breaking code (for example, impersonating IT support).

Cryptography and encryption

Cryptography is the writing of secret codes. Encryption converts a message to a coded format, and decryption is deciphering the encrypted message. Security also involves encrypting data before transmission so that it remains protected even if intercepted.

Public key encryption

Public key encryption uses open standards, meaning the algorithms are published, widely available, and discussed by experts and interested parties. The algorithm is not the secret; the key is what keeps information secret until the intended recipient decrypts it.

Securing the Internet: certificates and trust

The Internet is based on a “trust” model. Digital certificates can be purchased from Certificate Authorities (CAs) to identify trusted sites. CAs issue certificates that businesses, organizations, and individuals load onto their websites. These certificates verify to web browsers that encryption keys belong to the business, enabling secure online purchases and the sending and receiving of secure documents.

Example: A suspicious school email

If a student receives an email saying their account will be deleted unless they confirm their password:

• Likely threat: phishing.
• Impact: if they comply, an attacker may access school systems, private messages, or reset passwords elsewhere.
• Responsible response: don’t click; verify through official channels; report to IT.

Exam Focus

• Typical question patterns:
  • Identify a cybersecurity or safety risk in a scenario and describe its potential impact.
  • Explain how phishing or social engineering works at a high level.
  • Describe behaviors that reduce harm (strong authentication practices, verification, reporting).
• Common mistakes:
  • Mixing up privacy and security in explanations.
  • Treating threats as purely technical while ignoring the human factor.
  • Giving generic advice instead of naming a specific risk and a specific preventive action.

Economic, Social, and Environmental Impacts: Jobs, Communities, and the Planet

To analyze computing’s impact well, you need to look beyond individual users. Computing reshapes economies, social structures, and environmental systems.

Automation and the future of work

Automation uses technology to perform tasks with reduced human involvement, enabled by software, data analysis, and increasingly machine learning-based tools. Automation can increase productivity, reduce costs, and improve safety in dangerous jobs, but it can also displace workers in some roles, shift jobs toward new skill requirements, and increase inequality if benefits concentrate among technology owners. A nuanced view notes that new jobs appear (developers, data analysts, technicians), but transitions can be painful and uneven.

The gig economy and platform power

App-based work platforms can provide flexible income, but they can centralize power: platforms set rules and pay structures, workers may have limited transparency into how assignments are decided, and ratings/algorithms can strongly affect livelihoods. This connects to fairness and accountability when algorithms influence income.

Social interaction and mental health

Computing innovations reshape relationships. People can maintain long-distance connections and find communities, but they can also face harassment, social comparison, and overuse. Design choices—notifications, infinite scroll, and recommendations—shape behavior, so impact isn’t only about “self-control.”

Environmental impacts: energy and e-waste

Computing has environmental costs: resource extraction and manufacturing, energy consumption by data centers and networks, and e-waste (discarded electronics containing hazardous materials). Computing can also support environmental solutions through smart grids and energy optimization, climate modeling and monitoring, and remote collaboration that reduces some travel.

Example: Streaming video services

Streaming expands access and is convenient.

• Beneficial: affordable entertainment; global distribution for creators.
• Harmful: increased data traffic and energy use; potential consolidation of media power; privacy risks from viewing data.

Exam Focus

• Typical question patterns:
  • Explain one way computing changes jobs or economic opportunities.
  • Describe a societal impact of a widely used platform (communication, mental health, civic discourse).
  • Identify an environmental cost and a possible mitigation approach.
• Common mistakes:
  • Claiming “automation eliminates jobs” without acknowledging job transformation and creation.
  • Treating environmental impact as only “electricity use,” ignoring manufacturing and e-waste.
  • Giving opinion-based answers without tracing a cause-and-effect chain.

Putting It All Together: How to Write Strong Impact-of-Computing Explanations

AP CSP impact questions reward clear reasoning more than fancy vocabulary. You can consistently write strong responses by focusing on stakeholders, data flows, and trade-offs.

A practical reasoning method

When given an innovation (for example, a smart doorbell, translation app, or generative AI tool):

1. Describe what it does in 1–2 precise sentences (include data collection/processing if relevant).
2. Identify stakeholders (users, non-users nearby, companies, governments, workers, communities).
3. Give one beneficial effect and explain the mechanism (how the tech causes the benefit).
4. Give one harmful effect and explain the mechanism (how the tech causes harm).
5. If asked, propose mitigations (policy, design changes, user practices) and acknowledge limitations.

Mini worked response: smart home speakers

Prompt style: “Describe one beneficial and one harmful effect of smart home speakers.”

• What it is: Smart speakers accept voice commands, send audio to cloud services, and use speech recognition to respond.
• Beneficial effect (mechanism): They improve accessibility for users who have difficulty typing or using screens by enabling voice-controlled messaging, reminders, and home automation.
• Harmful effect (mechanism): They can reduce privacy because microphones may capture sensitive conversations; stored recordings or transcripts could be accessed in a breach or used for targeted advertising.

The key is tying the impact to how the system works (microphones, cloud processing, stored data), not just stating “privacy” or “convenience.”

Exam Focus

• Typical question patterns:
  • Write or choose an explanation that correctly ties an impact to a feature of the technology (data collection, algorithms, network effects).
  • Compare impacts across stakeholders (benefit for users, harm for bystanders).
  • Evaluate a proposed solution and identify what risk it reduces.
• Common mistakes:
  • Listing effects without explaining the causal mechanism.
  • Using absolute language (“always,” “never,” “completely safe”) instead of trade-off reasoning.
  • Forgetting stakeholders who didn’t choose to participate (bystanders in surveillance, communities affected by rerouted traffic, workers displaced by automation).