Should You Disclose That You Used AI to Write?
Writers, companies and students increasingly face a simple question with no simple answer: do readers need to be told that AI contributed to a piece of writing? Norms are still forming, and requirements differ by field, platform and country.
Arguments for disclosure
- Trust: readers who later discover undisclosed use may feel misled.
- Accountability: disclosure makes clear that a human is responsible for checking accuracy.
- Context: in education and research, the level of independent work matters.
- Rules: some publishers, schools, contests and platforms require it.
Arguments for a lighter approach
- Spell-checkers and grammar tools have long been used without notice, and the line between those and a generative assistant can blur.
- Over-disclosure of trivial help may clutter writing and invite misplaced suspicion.
- Authorship has always involved editors, ghostwriters and templates that go unmentioned in many contexts.
A way to think about it
Ask what a reasonable reader would want to know. The more AI shaped the substance of the work, such as the ideas, the wording of large sections or the images, and the more the reader relies on the author’s personal experience or expertise, the stronger the case for stating it.
| Use | Disclosure usually advisable? |
|---|---|
| Spelling and grammar checks | Rarely needed |
| Brainstorming ideas that you then developed | Often optional, check local rules |
| Drafting substantial passages | Generally yes |
| Fully generated piece with light edits | Yes |
| Academic or regulated submissions | Follow the stated policy |
How to disclose well
- Be specific: “An AI assistant helped with an initial outline and with tightening the language. The author checked all facts.”
- Place it where readers will see it, such as a note at the end or beginning.
- Avoid apologetic or boastful wording.
- Keep it consistent across your publications.
Legal and platform considerations
Some advertising and political content rules, and some platform policies, already address synthetic or machine-assisted material. These change, so check the current terms where you publish. Disclosure of sponsored content is a separate duty that continues to apply.
Examples of plain disclosure wording
- “This article was drafted with the help of an AI tool and edited and fact-checked by our staff.”
- “The author used an AI assistant to suggest headline options; all text was written by the author.”
- “Translations in this post were produced with machine assistance and reviewed by a fluent speaker.”
Simple, specific statements like these are easy to read and avoid both secrecy and overstatement. Adjust them to your actual process, and update them if your methods change.
Internal disclosure matters too
Disclosure is not only for the public. Colleagues, editors and clients may need to know how a piece was produced so that they can assess risks, apply review procedures and meet contractual obligations. Teams that agree on internal reporting, such as a line in the content log noting the tools used, avoid disputes and make it easy to answer questions later. Transparency within the team is the foundation for transparency outside it.
Policies will continue to change, so revisit them periodically.
Whatever you decide, avoid deceiving the audience about who is responsible. Even where no rule exists, honesty about process tends to protect credibility, and credibility is the one asset writers cannot generate on demand.