Lesson 3 of 4 · about 5 min
Guardrails for journalists
When you finish, you can use AI in reporting without exposing sources, publishing invented facts or breaking newsroom rules.
Stuck? Ask the AI tutorProtect sources and unpublished work
Never paste the name of a confidential source, a leaked document, unpublished story drafts, or private messages into a consumer AI tool. You may not control where that data is stored or who can access it. Source protection is a basic promise of the job, and in some places the law gives reporters some protection for their sources. Rules differ, so check yours.
Follow your newsroom's AI policy. If there is none, ask your editor, and in the meantime use invented examples or general descriptions. Remember that metadata and file names can reveal as much as the text itself.
Verify everything
AI can invent quotes, experts, studies, court rulings and statistics. It can also produce convincing fake images, audio and video, and you may be handed them as tips. Verify claims with primary sources, confirm identities, and check images and recordings with multiple methods and colleagues. Reverse image search and contacting the original source are standard steps.
Do not use AI detectors as proof. They are unreliable and can falsely accuse people. Treat them as one weak clue among many. If you cannot verify something, say so, or do not publish it.
Fairness, bias and harm
AI tools can repeat stereotypes and may describe communities unfairly. Read your copy for assumptions about race, gender, religion, disability, class and place. Be careful with crime reporting, where tool-written summaries may sound more certain than the facts allow.
Think about the person named in the story. A mistaken detail from a tool can harm a real life. Follow your outlet's code of ethics and any professional standards of your region, and keep a human editor responsible for sensitive decisions.
Disclosure and saying no
Be open about AI use according to your newsroom's policy. If AI helped with a transcript or translation, consider a short note when it matters to readers. Never publish AI-generated text or images as if they were reported or photographed by a person.
You can say no. If someone asks you to use a tool in a way that breaks a source's trust, or to publish an unverified viral claim, you can say, “I will run it when we have confirmed it.” Being right matters more than being first.
Practice now
- Write a short list of what you would never put into an AI tool in your newsroom, and what you could safely use as a substitute.
- Take a viral image or claim in your feed. Write down the verification steps you would take before reporting it, without using an AI detector.
Use any AI assistant you have. Do not paste private, client or patient information into it.
Check yourself
1. Why shouldn't you paste a leaked document into a consumer AI tool?
2. How should AI image detectors be treated?
3. A viral claim has no confirmation. What is the best choice?
Ask the AI tutor
Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.
AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.