News Summary Primary-Source Check
For the method in Before You Trust an AI News Summary, Check the Primary Source. Independently measured sourcing error rates make this a five-minute habit worth keeping, not a one-time exercise.
Not for
A claim that arrived as a forwarded post rather than a chatbot summary - use the viral claim source-trace worksheet instead.
Step 1: The AI summary as given
- Tool used:
- Search/browsing mode on? ☐ Yes ☐ No
- Summary text or key points:
Step 2: The three claims that matter most
| Claim | Type (number / quote / attribution / outcome) |
|---|---|
Step 3: Primary source found
- Original outlet/article:
- Link:
- Named in the AI summary? ☐ Yes ☐ No
Step 4: Side-by-side comparison
| Claim | Matches exactly | Matches with different wording | Not supported |
|---|---|---|---|
| ☐ | ☐ | ☐ | |
| ☐ | ☐ | ☐ | |
| ☐ | ☐ | ☐ |
Step 5: Opinion vs. fact check
- Anything the summary states as fact is actually the article’s own reporting, not a quoted opinion
- No editorializing or framing was added that isn’t in the original
Decision
☐ Safe to repeat/act on as summarized ☐ Repeat only with corrected/primary-source wording ☐ Do not repeat - could not confirm against a primary source
The two prompts to reuse
1. Extract the checkable claims
Here is an AI-generated summary of a news story: [paste it]. List the
three most specific, checkable factual claims in this summary -
prioritize numbers, quotes, and anything attributed to a named person
or organization.
2. Compare against the primary source
Here is the original article: [paste relevant section]. Here are the
claims from the AI summary: [paste them]. For each claim, tell me
whether the original article supports it exactly, supports it with
different wording or nuance, or does not support it at all. Do not
soften a mismatch - flag it plainly.