AI Ethics

Useful AI begins with a clearly defined purpose and depends on relevant, reliable, and sufficiently representative data. Poor, incomplete, or biased data can produce poor results even when the system is technically advanced. Machine colour, room size, and loudness normally do not determine whether an AI solution is appropriate or trustworthy.

AI systems may collect or process names, locations, photographs, voices, and other personal information. Privacy ensures that such information is collected, stored, used, and shared only for appropriate purposes and with proper permission. Therefore, option A is correct; the other options describe unrelated physical or technical activities.

Ethical rules guide the responsible design and use of AI. They help protect people’s rights, privacy, safety, dignity, and equal treatment. They also encourage checking data and results before important decisions are made. Thus option A is correct; changing colours or ignoring data accuracy is unrelated or unsafe.

An AI output may be incorrect, incomplete, biased, or unsuitable for the particular situation, even when it appears confident. Checking it against evidence, rules, or a reliable source helps detect errors before action is taken. Therefore A is correct; speed, storage, and automatic input changes are different issues.

AI may produce errors, unfair patterns, or results that do not fit the real context. Human supervision allows a person to review important outputs, question them, correct mistakes, and stop harmful actions. Therefore A is correct. Unchecked automation and removal of accountability are unsafe, while supervision does not primarily increase computer speed.

Fairness means that an AI system should not unjustly favour or disadvantage people, especially because of irrelevant personal characteristics. It requires attention to equal opportunity, appropriate treatment, and possible bias in data or outcomes. Helping only one person, hiding errors, or blocking someone without a valid reason is not fair decision-making.

Responsible AI use includes protecting personal information, limiting access, checking outputs, and avoiding harm. Option A reflects privacy protection. Option B violates consent and privacy, option C ignores the possibility of error, and option D deliberately causes harm. Therefore only A represents safe and responsible behaviour.

Safety in AI means designing, testing, and using a system so that foreseeable risks and harmful effects on people, property, or the environment are reduced. It also includes suitable safeguards and human review. Accepting unchecked decisions, exposing personal data, or spreading misinformation increases risk rather than providing safety.

A chatbot can produce useful answers, but it may also generate inaccurate, incomplete, outdated, or misunderstood information. Comparing its response with a textbook, reliable source, or teacher helps confirm correctness before use. Speed is not evidence of truth, and submitting copied AI text in an examination is not responsible learning.

AI can create a poster quickly, but it may insert an incorrect date, time, venue, spelling, image, or other detail. The student should review the complete poster, compare facts with the event information, correct errors, and then seek approval if needed. Sharing an unchecked poster can confuse participants and spread misinformation.

AI-generated career information may contain factual errors, outdated details, unsupported claims, or unnecessary personal data. The student should verify important facts with reliable sources, remove private information, and review the wording before publication. Options A and B ignore verification, while D exposes classmates’ data without a legitimate need or permission.

Responsible AI considers whether a system can cause harm, treat people unfairly, or misuse personal information. Therefore safety, fairness, and privacy are important principles, along with accountability and appropriate human oversight. Speed and cost may matter for design, but they alone cannot show that an AI system is responsible or ethical.

AI can make mistakes, reflect bias in its data, or give an answer that does not fit the real situation. Human supervision allows people to review outputs, ask for evidence, correct errors, and stop harmful actions. Keeping a machine off, ignoring outputs, or surrendering every decision does not provide responsible oversight or accountability.

A chatbot generates text from learned patterns and the information available to it; fluent wording is not proof that the content is true, current, or complete. Important facts should therefore be checked against reliable sources, especially for school, health, legal, or safety decisions. Chatbots do not possess human experiences simply because they use personal-sounding language, and systems may need updating.

AI can produce errors because of poor data, unusual inputs, bias, or limitations in its design. Human supervision allows people to review important outputs, notice harmful or unfair results, correct mistakes, and decide when the system should not be followed. It is not meant to hide errors, expose private information, or surrender every decision to a machine without judgment.

Transparency provides understandable information about what a system is intended to do, what data or factors it uses, and how its output should be interpreted. This can help users question errors, identify unfair treatment, and assign responsibility. Transparency is not the same as revealing every technical detail, but it is the opposite of hiding the basis of a decision or making it impossible to understand.

AI is a tool created, configured, deployed, and used by people. The relevant people and organisations remain responsible for choosing its purpose, checking its outputs, protecting affected users, and taking action on decisions. An AI system cannot independently understand moral or legal responsibility, and responsibility does not disappear merely because software was involved in producing an outcome.

Ethical AI use requires honesty about how AI is being used, protection of personal information, attention to safety, respect for people, and responsible checking of outputs. Obtaining data without permission can violate privacy, spreading misleading content can cause harm, and hiding known errors prevents accountability. Ethical practice aims to reduce harm and preserve trust rather than conceal problems.

An AI career app can offer useful possibilities, but it may not know a student’s complete interests, abilities, family situation, or opportunities. Option B is correct because the recommendation should be compared with personal goals and discussed with a trusted teacher or counsellor. AI should support, not replace, human judgement.

Reliable AI use depends on relevant, good-quality data and careful checking of the system’s results. Option D includes both safeguards. Verification can reveal errors, bias, or unsuitable recommendations before they cause harm. Unchecked outputs, biased training data, and an unclear purpose reduce trustworthiness rather than improve it.

Human review is especially important when an AI decision may affect hiring, loans, healthcare, education, safety, rights, or access to opportunities. Errors or bias in such settings can seriously harm people. Option A is correct; speed, internet availability, or a simple calculation does not by itself determine the level of oversight needed.

Training data should represent the relevant groups and situations fairly, with quality checks for missing or distorted information. Diverse and balanced data can reduce the chance that a model systematically favours or harms one group, so option B is correct. It does not remove all bias automatically, but it is an important safeguard.

Before using an AI system, the user should understand what it is intended to do, what data it uses, and what risks may arise. This supports transparency, privacy protection, and careful checking of results. Option D is correct. Unauthorised collection, hiding errors, and unclear purpose are unsafe and unethical practices.

A responsible user treats AI output as assistance rather than unquestionable truth. The user should understand its meaning, check important facts and sources, consider possible bias, protect personal information, and use the result appropriately. Option A includes these duties; the other choices encourage misinformation, privacy violations, or blind dependence.

A fluent or confident AI response is not proof that the information is correct, complete, or suitable for a particular person. Health decisions can have serious consequences, so the student should verify the information with a qualified professional or reliable medical source before acting. Option A follows safe, critical use of AI.