How to Use AI Right

AI assistants such as ChatGPT, Claude, and Gemini can write, analyze, summarize, code, and research in seconds.

Used well, AI helps you:

• Think more clearly
• Work faster on routine tasks
• Learn new subjects
• Explore more options
• Catch mistakes
• Create better first drafts

Used badly, it spreads errors, weakens your skills, and exposes private information.

AI multiplies what you bring to it. Bring clear thinking.

Before typing, decide:

• What outcome you need
• Who it is for
• What good looks like
• How you will use the answer

A clear goal turns a generic answer into a useful one.

AI cannot read your mind. Include:

• Background and situation
• Audience
• Constraints
• Examples of what you like
• Relevant files or data
• What you have already tried

Most bad AI answers are answers to bad questions.

Let AI produce options, drafts, and analysis. Keep decisions with people.

• “Give me three approaches with trade-offs.”
• “Draft this, and I will edit it.”
• “What am I missing?”
• “Argue against my plan.”

AI can state wrong information confidently. Check:

• Facts and figures
• Quotes and citations
• Dates and recent events
• Legal, medical, and financial claims
• Code before it runs anywhere important
• Calculations

Never assume confidence means correctness.

Ask AI to cite sources, then open them. Prefer:

• Primary sources
• Official documentation
• Recent publications
• Reputable outlets

A citation you have not opened is not a check.

Do not paste in:

• Passwords and API keys
• Government ID numbers
• Bank and card details
• Other people’s personal data without permission
• Confidential company information on personal accounts

Check each tool’s data settings, and use business plans for company work.

If AI does all your thinking, your skills fade. Protect them:

• Try first, then compare with AI
• Write your own outline before asking for a draft
• Ask AI to question you, not just answer you
• Explain the result in your own words
• Do some tasks without AI on purpose

Use AI to think harder, not to avoid thinking.

• “Explain this simply, then quiz me.”
• “Do not give me the answer. Give me hints.”
• “What is the most common misunderstanding here?”
• “Give me a practice problem.”

Learning with AI means doing more of the work yourself, with faster feedback.

Quick answers

A chat assistant

Current information

An assistant with web search or research mode

Your documents

Projects or source-grounded tools such as NotebookLM

Coding

Coding agents such as Claude Code or Codex

Repeated workflows

Skills, templates, or automations

• Point out what is wrong
• Ask for specific changes
• Request alternatives
• Ask what assumptions were made
• Start a new chat when the topic changes

Long, unfocused chats lose quality. Fresh context often helps.

• Follow your school’s or employer’s rules
• Disclose AI use where it is expected
• Never present AI output as personal experience
• Label realistic AI-generated images and video
• Respect copyright and other people’s work

Keep people in charge of:

• Final decisions
• Anything that affects others’ health, money, or safety
• Hiring and personal judgments
• Legal and financial commitments
• What gets published under your name

You are accountable for what you do with AI’s output.

• Asking vague questions
• Accepting the first answer
• Trusting citations without checking
• Pasting in sensitive data
• Letting AI make important decisions
• Using AI for everything, including what you should learn
• Ignoring tool limits and dates

Goal

Know what you want.

Context

Give the background.

Draft

Let AI create options.

Question

Challenge the answer.

Verify

Check what matters.

Decide

Keep judgment human.

Learn

Get better, not dependent.

I want to [goal] for [audience]. Here is the context: [background, constraints, examples]. Before answering, ask me any questions you need. Then give me a draft or options, explain your assumptions, list what I should verify, and tell me what you are unsure about.