Understand what AI can and cannot do, and learn the safety rules that protect you.
A general-purpose AI assistant is a very fast text-prediction system trained on large amounts of public text. It does not think, remember you across sessions by default, or have opinions the way a person does. It generates the most likely next words based on what you asked. That is useful for drafting, summarizing, and explaining — and it is also why it can sound confident when it is wrong.
AI is good at: rewriting messages, brainstorming lists, explaining unfamiliar terms, summarizing long text, and translating rough ideas into clearer language. It is weak at: fresh facts about your specific life or business, exact numbers, private data it cannot see, and anything requiring a real person's judgement — such as legal, medical, or financial advice.
Before you paste anything into an AI, ask three questions. One: would I be comfortable if this ended up public? Two: is any of this someone else's private information? Three: does the answer need to be verified before I act on it? If the answer to any of those is uncomfortable, remove the sensitive part, use a fictional example, or check a trusted source before you use the reply.
Treat the first reply as a draft, not a verdict. Skim it for anything that looks made up, especially names, statistics, dates, or quotes. Ask the AI to explain how it reached a conclusion, or to give sources. If it cannot, treat the specific claim as unverified.
A learner writes to ChatGPT: 'I run a small bakery. Rewrite this message to a customer who asked if we deliver on Sunday: Hi, we don't deliver Sundays. Sorry.' The AI returns a warmer version. The learner keeps the meaning, edits one line to match their own voice, and sends it. No customer name or phone number was shared.
You are a helpful writing assistant. Rewrite the message below so it stays honest, is warm and clear, and is no longer than 3 sentences. Audience: [describe customer or reader, e.g. a first-time customer]. Goal: [what should the reader do or feel next]. Keep the meaning. Do not invent facts. Do not add promises I did not make. Message: [paste your draft here]
Never include real customer names, phone numbers, addresses, or order numbers. Replace them with placeholders such as [customer name].
Open any free AI chat you have access to. Send the rewrite prompt above using a short, made-up message of your own (2–3 sentences). Read the reply, edit one line so it sounds like you, and save both the original and the edited version in a plain note.
Self-check only. Nothing here is scored. Tap a question to reveal the answer.
Write two sentences: one useful thing AI could help you with this week, and one task where you would not trust AI without checking.
This first lesson establishes the habit that protects the next 27: no real personal, customer, financial, medical, legal, or confidential business data goes into AI tools. Use fictional stand-ins.
Your progress is saved on this device only. Clearing browser data may remove it.
This is a Shipzo Academy proof-of-completion certificate. It is not a university degree, government qualification, accreditation, or regulated professional license. Completion does not guarantee employment, income, or acceptance into any Shipzo program. Tool availability, features, and free access may change and vary by country. Practice with examples — never paste real customer or personal data into AI tools.