Ideabro is an AI-first magazine, which means we use AI tools in our work and build AI features into the reading experience. It also means we owe you complete clarity about how we do it. This policy explains what our labels mean, what we will and won't use AI for, and how we check everything before it reaches you.
Our three labels
Every article on Ideabro carries one of three labels, shown near the byline.
| Label | What it means | Example |
|---|---|---|
| Human-written | A person researched, reported and wrote the piece. AI may have been used for minor tasks like spell-check or transcription, but not for generating the text. | An opinion column, a reported interview |
| AI-assisted | A person wrote the piece, but AI tools meaningfully contributed — for example, by summarizing research, suggesting structure, tightening drafts or helping build a table. The author is responsible for every word. | A guide where AI helped organize test notes |
| AI-generated | The text was substantially drafted by an AI system, then fact-checked and edited by a human editor before publication. | A glossary drafted by Ideabro Machine |
When a piece is labelled AI-assisted or AI-generated, we add a short note explaining how AI was used.
What we use AI for
Our writers and editors may use AI tools to:
- Transcribe interviews and recordings.
- Summarize long documents, research papers and public filings as a starting point for reading them.
- Brainstorm angles, headlines and questions.
- Suggest structure or edits to a draft.
- Check grammar, consistency and clarity.
- Translate material for research purposes, with key passages confirmed by a fluent speaker.
- Draft clearly labelled explainers, glossaries and roundups through Ideabro Machine.
What we don't use AI for
- Inventing quotes or sources. Every quote in our journalism comes from a real conversation or a verifiable document. When a piece uses an illustrative scenario or a composite character, we say so plainly in the piece.
- Fabricating reporting. AI does not conduct interviews or claim to have witnessed events.
- Generating images of real people or events and presenting them as authentic.
- Writing opinion pieces. Opinions on Ideabro belong to named human authors.
- Publishing without review. No AI-generated text reaches readers without a human editor's approval.
How we review AI-involved work
Any piece involving AI goes through the same editing as everything else, plus extra checks:
- Fact-check every specific claim — names, numbers, dates, quotes, technical details — against primary or authoritative sources.
- Check for invented references. We confirm that every cited paper, product, person and organization exists and says what we claim.
- Read for sameness. AI drafts can drift toward generic phrasing. Editors rewrite for clarity and voice.
- Confirm the label. The editor responsible for publication confirms the disclosure label is accurate.
AI features for readers
Ideabro offers AI-powered features, including article summaries, plain-language explanations, suggested questions, the Ask the Archive tool and the Idea Lab. These are generated automatically and labelled as AI output.
They can make mistakes. Summaries may miss nuance; answers may misread a story. When accuracy matters, read the original article, which is always linked. Our AI answers are designed to cite the Ideabro stories they draw from, so you can check them.
If an AI feature gives you an answer that's wrong, misleading or inappropriate, please tell us at corrections@ideabro.com. We review these reports and use them to improve the features.
Privacy and AI
When you use our AI features, your questions may be processed by a third-party AI provider to generate an answer. We don't use your questions to identify you, we don't sell them, and we ask our providers not to use them to train their models where that option is available. Don't enter sensitive personal information into AI features. See our Privacy Policy for details.
Corrections
If we get something wrong — whether a human or an AI made the error — we fix it promptly and transparently, following our corrections policy. We do not quietly edit published facts.
This policy will change
AI tools and the norms around them are evolving quickly. We'll update this policy as our practices change and note the date of each revision here. If you have questions or suggestions, write to hello@ideabro.com. We'd genuinely like to hear them.
Last updated · Questions? hello@ideabro.com