Using AI writing tools responsibly: what a human review catches
AI drafts are fluent but not always right. Here is what to check before you publish.
AI tools write fluent text quickly. Fluent is not the same as accurate, original or right for your reader. Before you publish anything an AI helped write, check it properly.
What to check
Facts and figures
AI can state wrong facts confidently. Check every number, date, name and claim against a reliable source.
Sources
AI tools sometimes invent references, quotations and links. Open every source and confirm it says what the text claims.
Tone and voice
AI drafts often sound generic. Make sure the text sounds like your organisation and suits your reader.
Purpose
Check the text does what you need: answers the question, makes the case or gives the instruction clearly.
Signs a draft needs work
- the same point made three ways;
- vague, impressive-sounding words with little meaning, such as "leverage", "robust" or "seamless";
- every paragraph the same length and shape;
- confident claims with no evidence;
- headings and lists where plain sentences would read better.
Rules to follow
- Check your organisation's or institution's policy on AI use. Many universities do not allow AI-written assessed work.
- Do not paste confidential information into tools that may store or learn from it.
- Take responsibility for what you publish, however it was drafted.
How we handle AI
Our Human review of AI-generated writing service gives your AI-assisted text a professional edit for accuracy, clarity and tone. And if you would rather your document never reaches AI at all, choose Human editing: Human orders are never sent to generative AI.
Want a second pair of eyes?
Upload your document for an instant, itemised quote. You see the full price before you pay.