Introduction to AI

Artificial Intelligence refers broadly to methods that enable computer systems to perform tasks such as recognising patterns, understanding language, making predictions, or supporting decisions. It is used in many software systems and devices, not only robots or internet searches. AI depends on data, instructions, and design choices, so it can make mistakes and does not always produce a correct answer independently.

AI learns from data and follows the objectives, instructions, and design provided by people. If the training data are inaccurate, incomplete, or biased, the system may reproduce those problems in its outputs. AI can also fail on situations unlike its training examples. Therefore it is not guaranteed to be always correct, fair, or better than human judgment, and human review remains important.

In AI and machine learning, a feature is a measurable property or piece of information used to describe an example and help distinguish it from others. A useful feature has a meaningful relationship with the task, such as shape or colour when identifying an object. Unrelated information, deleting data, or the machine’s appearance does not by itself help the system identify the target.

Artificial Intelligence can analyse weather records, soil conditions, crop growth, and past farm data to provide useful advice about sowing, irrigation, fertiliser use, and possible pest attacks. Therefore, option A is correct. The other choices describe unsafe or unhelpful practices that ignore evidence and differences between fields.

Automatically sorting emails into spam, priority, and general folders is an example of AI-assisted time saving. An AI system can examine patterns such as words, senders, and previous choices, then place messages in suitable categories. This reduces repeated manual work while still allowing the user to review important messages. Reading every email by hand takes more time, deleting all messages is unsafe, and taking no action is not automation.

A music application that recommends songs from a person’s listening history is using an AI-based recommendation process. It can identify patterns in earlier choices and predict what the listener may enjoy, although such predictions are not always perfect. A printed calendar, a wall clock, and a handwritten shopping list provide fixed information or record information; they do not analyse behaviour to generate personalised suggestions. Thus, option A is correct.

A balanced view recognises both the usefulness and the limitations of AI. AI can help with tasks such as translation, recommendations, and data analysis, but its output may contain errors or reflect unfair patterns in the data used to develop it. People should therefore protect privacy, consider fairness and safety, and check important results. Saying that AI is always wrong, only for entertainment, or safe to trust without checking is not balanced.

Automation allows software or machines to perform defined steps with little repeated human intervention. In AI-supported work, this can save time on routine activities such as sorting records, entering predictable data, or preparing a first draft of a report. Automation does not guarantee perfect quality, and it should not be used to hide mistakes; outputs still need suitable supervision. Thus, saving time in repetitive tasks is the valid benefit in option D.

A facial-recognition system can compare facial features with authorised records and support identity checks before access is granted. Thus, option B is an example of AI-assisted security. It should still be used lawfully and with privacy safeguards. Unchecked entry, public data sharing, or ignoring failed verification increases risk.

AI is not limited to physical robots. It also appears in translation, recommendation systems, education tools, healthcare support, agriculture, and many software applications. Therefore, option D is the misconception. The other statements are reasonable: AI can find patterns in data, assist many tasks, and still benefit from human supervision.

A song recommendation system analyses listening history and preference patterns to suggest content that may interest a particular user. This is a personalised AI application in entertainment, so option B is correct. The other activities are ordinary manual or simple counting tasks and do not involve such data-based recommendation.

An AI learning tool can examine a student’s answers, errors, and progress and then recommend practice suited to that learner’s needs. Option C is therefore personalised support. Giving everyone identical work, hiding progress, or closing the classroom does not adapt learning to an individual student.

AI can benefit society when its goals are useful, its data and methods are handled fairly, privacy and safety are protected, and important results receive suitable human oversight. Option A includes these conditions. Unauthorised data sharing, unverified misinformation, and blind dependence can damage rights, trust, and safety.

AI refers to systems designed to perform tasks that commonly require human-like abilities, such as recognising patterns, processing language, making predictions, or supporting decisions. These systems depend on data, rules, or training and can make mistakes. Thus, option A is the most appropriate; the other options make false absolute claims.

In many AI and machine-learning systems, the machine is given examples represented as data and learns patterns or relationships from them. For instance, labelled images can help a system distinguish categories. Physical details such as the machine’s colour, room height, or chair design are irrelevant unless they are deliberately included as meaningful data for a particular task.

AI works more effectively when the problem and objective are clearly defined and the data is accurate, relevant, and sufficient. These conditions help a model learn meaningful patterns and allow people to evaluate its results. Vague goals, poor data, or blind acceptance of outputs can produce unreliable or unfair decisions, so human checking remains important.

During training, an AI model learns relationships or patterns from examples. Afterward, it can apply those learned patterns to new inputs and estimate a category or likely outcome. This is generalisation, not simple memorisation. However, performance depends on the quality and relevance of training data, so training does not guarantee perfect decisions in every unfamiliar situation.

Prediction uses patterns in earlier data to estimate a future result. Forecasting tomorrow’s weather from past weather observations is therefore a prediction task. Identifying a cat is image classification, grouping similar customers is clustering, and preparing a summary is a language-generation or text-processing task.

AI systems learn from data and instructions, so poor-quality, incomplete, or biased data can lead to inaccurate or unfair outputs. More data does not guarantee perfect decisions in every situation. AI may also miss its own errors, and new tasks generally require suitable data, design, or human guidance.

The statement is false because AI is used in many software-based systems that are not robots. Examples include voice assistants, spam filters, recommendation systems, translation tools, and learning platforms. The word “only” makes the statement incorrect; robots are just one possible form in which AI may be used.