Unit 4: AI Tools You Can Use

SSC200 — Fundamentals Of Artificial Intelligence 10 min read

I. Orientation: Practical Artificial Intelligence

Artificial intelligence (AI) refers to computer systems that perform tasks commonly associated with human intelligence, including understanding language, generating content, recognizing patterns, making predictions, and supporting decisions. Modern generative AI tools use machine-learning models trained on large collections of text, images, audio, code, or data. They produce outputs by identifying patterns in a user’s input and generating a suitable response; they do not automatically guarantee truth, originality, fairness, or professional quality.

Human direction: The user supplies a goal, context, instructions, and source material; the AI provides assistance rather than taking complete responsibility.
Prompt dependence: Output quality depends on the clarity of the prompt, the information provided, and the tool’s capabilities.
Probabilistic output: AI generates likely responses rather than retrieving guaranteed facts in every case.
Verification requirement: Names, dates, statistics, quotations, calculations, citations, and interpretations must be checked against reliable sources.
Privacy and security: Confidential records, passwords, personal identifiers, and sensitive organizational data should not be entered into tools without authorization.
Human accountability: The person submitting a CV, report, design, or analysis remains responsible for accuracy, ethical use, and final decisions.
Responsible originality: AI-generated material should be reviewed for plagiarism, bias, copyright concerns, misleading claims, and excessive similarity to existing work.

II. Introduction to Simple AI Tools — Basic Functions and Selection

A. Introduction to simple AI tools

Simple AI tools are accessible applications that assist with language, images, office work, research, organization, and data interpretation. Their purpose is to reduce routine effort while preserving human judgment.

Conversational assistants: Chat-based tools can explain concepts, draft text, reorganize notes, create outlines, and suggest alternatives from a natural-language prompt.
Text-processing tools: Grammar checkers, translators, paraphrasers, and transcription systems transform existing text or speech. A transcription tool, for example, converts recorded speech into editable text.
Image and design tools: Template-based design platforms can suggest layouts, color combinations, illustrations, presentations, or social-media graphics from a description.
Productivity tools: AI features in word processors, email systems, and spreadsheets can summarize meetings, suggest replies, classify messages, or generate formulas.
Tool selection: Choose a tool according to the task, required accuracy, privacy level, output format, cost, and need for specialist knowledge.
- Low-risk use: Brainstorming titles or reorganizing personal notes.
- Higher-risk use: Medical, legal, financial, employment, or academic decisions requiring expert review.
Effective prompting: A useful prompt identifies the role, task, context, constraints, audience, and desired format. For example: “Create a 150-word formal introduction for first-year students using the following topic and three supplied facts.”

III. AI for Summarizing — Reducing Information Efficiently

A. AI for summarizing

AI summarization condenses a longer source into fewer words while attempting to preserve its central ideas, structure, evidence, and conclusions. It is most reliable when the source is supplied directly and the required length or purpose is clearly specified.

Extractive summarization: The tool selects or slightly shortens sentences from the source. This can preserve wording but may produce a disconnected summary.
Abstractive summarization: The tool restates ideas in new language. It is smoother but may introduce inaccurate details or meanings not present in the source.
Prompt controls: Specify word count, audience, tone, and emphasis, such as “Summarize the argument in 100 words and retain all numerical findings.”
Source fidelity: Compare the summary with the original, especially for negations, qualifications, dates, percentages, and cause-and-effect relationships.
- Risky change: “The study found no significant effect” becoming “The study found an effect.”
- Risky omission: Removing the limitation that a result came from only 40 participants.
Useful formats: AI can produce bullet points, abstracts, executive summaries, tables of claims and evidence, or a section-by-section outline.
Academic integrity: A summary must represent the original author fairly and should not replace required reading or appropriate citation.
Worked example: For a 20-page report, request “five key findings, two limitations, and the conclusion, with page numbers.” Check each item against the relevant page before using it.

IV. AI for Writing — Supporting the Writing Process

A. AI for writing

AI writing tools support planning, drafting, revising, editing, and adapting communication for different audiences. They are most useful as collaborative assistants rather than invisible replacements for the writer’s thinking.

Planning: Ask for a thesis statement, outline, paragraph sequence, or list of evidence needed for a topic.
Drafting: Provide the purpose, audience, tone, length, and facts. A prompt such as “Draft a formal 200-word email requesting a meeting; do not invent dates” sets clear boundaries.
Revision: AI can identify repetition, weak transitions, unclear pronouns, excessive passive voice, or inconsistent terminology.
Style adaptation: The same content can be rewritten as a formal report paragraph, a public announcement, or an accessible explanation without changing verified facts.
Fact control: Supply a fact sheet and instruct the tool to use only that material; independently verify any added claim.
Human voice: The writer should decide the argument, examples, interpretation, and final wording. Automatically generated prose can sound generic or fail to reflect personal experience.
Language quality: Grammar correction is useful, but a technically correct sentence may still be unsuitable, culturally insensitive, or logically weak.
Integrity and disclosure: Follow institutional rules concerning permitted AI assistance, acknowledge substantial assistance when required, and avoid submitting generated text as wholly independent work.

V. AI for Designing — Creating Visual and Presentation Materials

A. AI for designing

AI design tools generate or refine visual materials such as posters, slides, logos, illustrations, layouts, and social-media content from written instructions or selected templates. Their value lies in rapid exploration of alternatives.

Design brief: State the audience, purpose, dimensions, platform, visual style, color preferences, text, and required elements before generating a design.
Layout generation: AI may arrange headings, images, icons, and whitespace, but the user must check hierarchy, alignment, readability, and balance.
Image generation: A detailed prompt can specify subject, setting, composition, lighting, medium, and aspect ratio. Generated images may contain distorted text, unrealistic objects, or unintended stereotypes.
Brand consistency: Use approved fonts, colors, logos, and tone; an attractive output is unsuitable if it violates an organization’s identity guidelines.
Accessibility: Check color contrast, font size, alternative text, captions, and whether information is communicated only through color.
Copyright and likeness: Verify licensing conditions for generated or source images and avoid using a recognizable person’s likeness without appropriate permission.
Presentation design: AI can turn an outline into slides, but each slide should communicate one clear idea rather than display dense AI-generated paragraphs.
Human review: Inspect every visual at its final size. A layout that appears acceptable in an editor may become unreadable on a mobile screen or projector.

VI. AI for Analyzing Data — Finding Patterns Responsibly

A. AI for analyzing data

AI-assisted data analysis helps users clean, classify, visualize, summarize, and model datasets. It can accelerate exploration, but reliable analysis requires correct data, suitable methods, and interpretation grounded in context.

Data preparation: Check column names, units, missing values, duplicate records, impossible values, and inconsistent categories before analysis.
- Example: “Male,” “male,” and “M” may represent one category but must be standardized deliberately.
Descriptive analysis: AI can calculate totals, averages, medians, percentages, distributions, and trends. Each statistic must match the question and variable type.
Visualization: Request a chart appropriate to the data: a line chart for change over time, a bar chart for category comparison, or a scatter plot for relationships between numerical variables.
Pattern detection: Machine-learning tools may identify correlations or clusters, but correlation does not prove causation.
Formula and code assistance: AI can generate spreadsheet formulas or code, but test them on known values and inspect filters, ranges, assumptions, and units.

PYTHON

average = sum(scores) / len(scores)

Here, scores is a list of numerical observations and average is their arithmetic mean; an empty list would cause an error and must be handled.

Bias and representativeness: A model trained on incomplete or historically biased data can produce systematically unfair results.
Interpretation: Report the dataset size, collection method, time period, missing-data treatment, method used, and important limitations.
Privacy: Remove unnecessary identifiers and use approved systems when handling employee, customer, student, or patient data.

VII. Usage of AI to create CV — Presenting Qualifications Clearly

A. Usage of AI to create CV

AI can help create a CV by organizing experience, improving wording, matching relevant keywords to a vacancy, and producing a consistent layout. It must not invent qualifications, achievements, employers, dates, or skills.

Information collection: Prepare verified details for education, employment, projects, certifications, technical skills, achievements, contact information, and references.
Job alignment: Paste the genuine job description and ask the tool to identify relevant keywords, such as “data visualization,” “customer service,” or “project coordination.”
Achievement-focused wording: Convert duties into evidence-based statements using action, task, and result.
- Weak: “Responsible for social-media posts.”
- Stronger: “Scheduled weekly social-media posts and monitored engagement using platform analytics.”
Quantification: Add accurate numbers where available, such as “processed 120 invoices per month” or “reduced response time by 15%.”
Structure: A clear CV commonly includes a professional profile, skills, experience, education, and selected achievements in reverse chronological order.
Applicant-tracking systems: Use standard headings, readable fonts, and job-relevant terminology; avoid placing essential information only in images, icons, headers, or complex tables.
Truth and ownership: Review every sentence and ensure that the applicant can explain and demonstrate each listed skill during recruitment.
Privacy and formatting: Limit sensitive personal information, check contact details, export to the requested file type, and inspect the final document for spacing and page breaks.
Appropriate prompt: “Rewrite these verified duties into four concise achievement bullets for a laboratory assistant role; do not add responsibilities or numbers.”

VIII. Usage of AI in the Process of Report Writing — From Evidence to Final Document

A. Usage of AI in the process of report writing

AI can support the report-writing process from planning through editing, but the report’s research question, evidence, reasoning, and conclusions must remain controlled by the writer.

Planning: Use AI to turn the assignment brief into a structure containing title, introduction, methodology, findings, discussion, conclusion, and recommendations where appropriate.
Research organization: Provide collected notes and ask for a table separating claims, supporting evidence, source details, and unresolved issues. Do not treat invented citations as sources.
Introduction: AI can help explain context, define the problem, identify objectives, and formulate a scope statement based on supplied information.
Methodology: Describe what was actually done: participants or data source, instruments, procedure, dates, and analysis method. AI must not fabricate a method after the research is complete.
Findings: Use verified results, tables, and figures. Keep findings distinct from interpretation; for example, “sales increased 12%” is a result, while “the campaign caused the increase” is an interpretation requiring evidence.
Discussion: Ask AI to identify relationships between findings and objectives, alternative explanations, limitations, and implications. Evaluate each suggestion against the data.
Editing: AI can improve headings, transitions, grammar, concision, consistency of tense, and formatting while preserving the writer’s meaning.
Citation control: Insert only sources actually consulted and verify author, title, year, page, URL, and citation style manually.
Confidentiality: An unpublished business report or identifiable research data should be processed only through an authorized tool.
Final quality check: Confirm that every objective is addressed, every major claim has evidence, tables match the text, recommendations follow from findings, and the conclusion does not introduce new unsupported information.