Report · August 5, 2026

We fact-checked 5,500 claims from AI-written news. 30% weren't supported by the sources.

1 in 3

factual claims had no support in the sources the system was given. Not wild hallucination — plausible context filled in from model memory.

Sample 1,140 articles / 5,500 individual claims
Window 2026-05-03 → 2026-08-04
Source systems operated by weranai — method in Methodology

We run an autonomous editorial pipeline: a set of agents that research, write, and publish news articles without a human in the loop. Every article passes through an automated fact-checker before publication, which reads each factual claim and tries to trace it back to the source material the writing agent was given.

Over three months that check ran on 1,140 articles and examined 5,500 individual claims. This is what it found.

Claim-level results

The unit that matters is the claim, not the article. An article can read perfectly and still contain a figure nobody can trace.

Verified 65.8%Flagged 29.7%Too recent 2.7%Corrected 1.8%
All 5,500 claims examined, by verdict.
VerdictClaimsShareMeaning
VERIFIED3,62165.8%Supported by the source packet
FLAGGED1,63529.7%No support found in the sources
RECENT1462.7%Too recent to verify at write time
CORRECTED981.8%Rewritten to match the source

Nearly one in three factual claims had no backing in the material the system had in front of it. Reading the flagged examples, almost none are absurd. They are plausible: a specific exchange rate, a year for a past event, a job title. The model fills gaps with things that sound right, and the gaps are invisible unless you check claim by claim.

A representative flag. An article stated that a currency “was trading above 163 yen last week — a 40-year high.” The sources confirmed a sharp move and an intervention, but contained neither the price level nor the 40-year framing. Both were supplied by the model.

Article-level results

A single unsupported claim is enough to hold an article back, so the article-level pass rate is much harsher than the claim-level one.

StatusArticlesShareMeaning
NEEDS_REVIEW83373.1%At least one unsupported claim
PASS26923.6%Every claim traced to a source
MINOR_ISSUES282.5%Small corrections applied
PENDING100.9%Check did not complete

Under 23.6% of articles came through with every claim traced. Put differently: left unattended, roughly three out of four AI-written news articles would have published at least one figure nobody could source.

Fabrication is not evenly distributed

The clearest finding is that the failure rate depends on the subject. The same models, the same prompts, the same pipeline — very different results by topic.

TopicCleanTotalPass rate 
Technology9832030.6%
Space279927.3%
AI5220525.4%
Markets/Investing5225120.7%
World politics4026515.1%

World politics fails roughly twice as often as technology. Our reading: topics with dense primary documentation — product specs, filings, release notes — give the writer something concrete to anchor to. Geopolitical coverage rewards synthesis and context, which is exactly where a model reaches for memory instead of sources.

If that holds generally, it implies the risk of using generative systems for reporting is not uniform. It is highest precisely where verification is hardest and the cost of being wrong is greatest.

What this does and does not show

  • The checker only ran on news articles. Procedural content written against vendor documentation was never part of this sample, and we would expect it to behave differently.
  • “Unsupported” means absent from the source packet given to the writer. Some flagged claims may well be true in the world — they simply were not traceable at write time, which is the standard that matters before publishing.
  • The fact-checker is itself a model. It has its own error rate, which we have not yet measured. That measurement is on the list.

Full method, definitions, and the exact status vocabulary are in Methodology.