Fact-Checking TikTok Claims With AI in Under 60 Seconds

Most viral 'science facts' on TikTok are wrong, exaggerated, or missing context. AI can help you check fast.

What to actually do

• Step 1: write down the specific claim in one sentence
• Step 2: ask Perplexity for sources for and against
• Step 3: check the original study, not the headline

The big idea: AI makes fact-checking 100x faster. The skill is knowing what to actually check.

Lots of older family members share fake news, AI-generated images, or misleading 'facts' on Facebook. You can be the family fact-checker by using AI quickly to verify before reacting.

Some examples

• 'Is this story (link) real? When did this actually happen?'
• 'Is this image real, or AI-generated?'
• 'Did this politician really say this exact quote?'
• 'Is this medical claim supported by research?'

Try it!

Next time you see something on social media that feels off, fact-check it with AI. Did your gut feeling match what you found? Practice makes you sharper.

Most essay errors aren't grammar — they're confidently wrong claims about dates, statistics, and quotes. AI is much better at flagging those than at writing the essay itself, because flagging requires comparison to known facts, not generation.

Some examples

• Prompt: 'Read my essay and list every factual claim — dates, numbers, quotes, names. For each, rate confidence high/medium/low and tell me which to verify.'
• Don't ask AI to verify the claim — it'll hallucinate. Ask it to flag the claim, then verify yourself with Wikipedia (footnotes), the original source, or a reputable encyclopedia.
• Quotes are the easiest to get wrong: AI catches when a quote attributed to Einstein is actually from Mark Twain or made up entirely.
• Numbers without sources ('40% of teens use AI daily') are almost always wrong; AI flags them so you can find a real cite or cut the claim.

Try it!

Take an essay you wrote recently. Paste it into Claude with: 'Extract every specific factual claim and put it in a numbered list. Mark which ones I should verify before submitting.' Then verify the top 3 with real sources. Watch how many were wrong.

Ask ChatGPT for a Lincoln quote and it might invent one that sounds perfect — and never existed. Same for sources, page numbers, even Supreme Court cases. Teachers catch these instantly because they search the source and find nothing. Always verify every quote and source before submitting.

Some examples

• A NY lawyer was sanctioned for citing AI-invented court cases.
• ChatGPT invents URLs that look real but go nowhere.
• Cross-check quotes on AmericanRhetoric.com or Wikiquote.
• Ask the AI for the source, then Google the source separately.

Try it!

Ask AI for a famous quote about freedom. Then Google the exact words. About 1 in 4 will be wrong or partly wrong — see for yourself.