1. Trusting answers without checking them
AI assistants can produce text that sounds right and is not. This includes made-up facts, wrong numbers and links or sources that do not exist. Official help pages for the major assistants say the same: results can be incomplete, outdated or incorrect.
Fix it. Ask for sources, open them, and confirm anything that matters: prices, dates, legal rules, names, math. If it cannot give a source, treat the claim as unverified.
For each factual claim in your last answer, give me the source (a link) and say how confident you are. Mark anything you are unsure about, and list what I should verify myself before using this.
2. Pasting private or sensitive data
Customer lists, passwords, bank details, ID numbers, health information and contracts do not belong in a chat box by default. Depending on the product, plan and settings, conversations may be stored, reviewed, or used to improve services. The rules differ from tool to tool and can change.
Fix it. Remove or replace names, numbers and identifiers before you paste ("Client A", "$X"). Read the privacy and data controls of the plan you use, and follow your employer's policy if you have one.
3. Giving vague instructions
"Write an email" gets you a generic email. The assistant cannot see your situation unless you describe it.
Fix it. Say who the reader is, what you want them to do, the tone, the length, and give an example if you have one. Then ask it to ask you questions if anything is unclear.
I want you to help me with [TASK]. Audience: [WHO]. Goal: [WHAT SHOULD HAPPEN]. Tone: [TONE]. Length: [LENGTH]. Before you start, ask me up to 5 questions if you need more context.
4. Sending AI output without human review
A draft is a draft. Sending it as is can mean wrong details, an odd tone, or promises you did not mean to make. For anything touching legal, tax, medical, safety or financial decisions, AI is a helper for questions to ask a professional, not a replacement for one.
Fix it. Read everything out loud once. Check names, dates and amounts. Put your own voice back in. Ask a qualified person when the stakes are high.
5. Assuming it knows what is current or specific to you
An assistant does not automatically know your business, your latest prices, or what happened this morning. Without web search or your own notes, it relies on what it learned in training, which has a cut-off date.
Fix it. For current topics, turn on web search (if your plan offers it) and ask for dated sources. For your own situation, paste the relevant (non-private) facts into the chat so it works from them.
Quick checklist
- Did I remove private data before pasting?
- Did I check the important facts against a real source?
- Did I give clear context, audience and goal?
- Did a human (me) read the final text?
Want a workflow that builds the checking in? See how to use Grok, ChatGPT and Claude together.
Grok, ChatGPT and Claude are products of their respective owners. AIBA Technologies LLC is not affiliated with or endorsed by them. AI tools change often, so features and limits vary by plan and over time. Check each product's own help pages for the latest details. This guide is general information, not legal, tax or financial advice.