Research and fact-checking with AI

In this lesson you will research one real question with AI, check every important claim in the answer, and fact-check a viral message. You will learn a repeatable five-step workflow you can use for assignments, projects, interview preparation and news.

What you need: a browser and free accounts for two or three of these: Perplexity, Gemini, ChatGPT, Microsoft Copilot or Google Search (which shows AI Overviews and has an AI Mode). Read *Hallucinations: why AI is confidently wrong* first if you haven't; this lesson is the practical follow-up.

A plain chatbot answer comes from what the model learned in training. It can be out of date and it has no sources to show you. An AI search answer works differently: the tool first searches the web, then writes an answer from the pages it found, with links. That makes it more current and checkable, but not automatically correct. The model can misread a page, mix up two sources or cite a page that doesn't actually say what the answer claims.

Research this question: [your precise question]

Context: I need this for [assignment / project / decision], and I'm in [country/region].

Format:
1. A short answer (3-4 sentences).
2. A table of the key claims with columns: claim | source name | link | date of source | fact or opinion.
3. Anything that is disputed or has changed recently.
4. What you could not find a reliable source for.

Prefer official and primary sources (government sites, regulators, company documentation, research papers) over blogs.

What to expect: The claims table turns a paragraph you can't easily check into a list you can tick off one by one. Point 4 invites the tool to admit gaps instead of filling them.

Professional fact-checkers use lateral reading: instead of judging a page by how professional it looks, they open new tabs and check what other reliable sources say about the claim and about the website itself. A well-known version of this is the SIFT method, created by researcher Mike Caulfield:

Stop: notice when a claim triggers a strong reaction or seems too neat. Investigate the source: who published this, and are they in a position to know? Find better coverage: do other reliable sources report the same thing? Trace to the original: follow the claim back to where it started, such as the actual study, the government notification or the full video.

When an AI tool lists research papers, some may not exist: the title, authors and journal look real, but the paper was never written. This is one of the most common hallucinations. Before you use any reference in an assignment or project report, look it up yourself.

Forwarded messages about new rules, fake job offers, health cures and "breaking news" spread fast. AI can help you check them, but a chatbot alone should not be your final word, because it may not know about events after its training or may repeat the claim back to you.

Use AI to break the message into checkable claims and suggest search terms. Then check those claims with AI search, official sources and fact-checking sites. In India, the PIB Fact Check unit checks claims about the central government, and independent organisations such as BOOM and Alt News publish fact-checks. Google Fact Check Explorer (toolbox.google.com/factcheck/explorer) searches fact-checks from many publishers at once.

Here is a message I received. Don't tell me whether it's true yet.

---
[paste the message]
---

1. List each separate factual claim in it.
2. For each claim, say who would be the official or original source to confirm it (e.g. a ministry, RBI, a hospital, the company).
3. Suggest 2 search queries for each claim.
4. Point out any warning signs (urgency, no source, requests for money or personal details, links to unofficial sites).

What to expect: By asking it not to give a verdict, you get a checking plan instead of a confident yes/no that may be wrong. Never click links or share personal details from a suspicious message while checking it.

AI answers tend to give one tidy answer. Real questions often have trade-offs, disagreements or different answers in different places. Ask for them directly, especially before decisions such as choosing a course, a laptop or a technology for your project.

Here is a conclusion I'm leaning towards: [your conclusion].

Give me the strongest evidence-based arguments against it, the situations where it would be wrong, and what information would change the answer. Cite sources where you can. Be direct; I want to stress-test it, not be reassured.

What to expect: A useful answer names specific conditions ("if you need X, this fails") and sources, not vague "it depends" statements.