How AI chooses, reviews and rates claims.

Every claim on this site is selected, evaluated, rated and described by an automated AI pipeline. There is no human editor reviewing each verdict before publication. We chose this model for transparency, scale and consistency — the same evidence rules are applied to every claim regardless of speaker.

AI systems can make mistakes. They can misread a statistic, miss context or over-weight a secondary source. That is why every rating is published with the full list of sources the AI relied on — so any reader can audit the reasoning and challenge it via the corrections page.

The AI reviews specific, verifiable factual claims made publicly — for example in speeches, press releases, parliamentary statements, recorded interviews and official social channels. Eligible claims are those a reader could in principle check against documented evidence.

• Opinions, value judgements and predictions about the future.
• Sarcasm, satire and obvious rhetorical flourishes.
• Vague claims that are not specific enough to be tested.
• Personal characterisations or attacks on individuals.

The AI is instructed to prefer primary sources: official statistics, original documents, raw datasets, legislative records and verifiable recordings. Secondary reporting may be included for context but should not be the sole basis for a verdict where primary evidence is reasonably available.

Claims are placed on a six-level editorial scale. The written explanation does the heavy lifting; the meter position is a reading aid.

• Accurate

  The claim is fully supported by credible primary evidence with no material omissions.

• Mostly Accurate

  Broadly supported by evidence but contains minor caveats or missing detail.

• Needs Context

  Factually correct in isolation but omits context that materially changes the meaning.

• Misleading

  Selectively combines or frames facts in a way that leads readers to a false conclusion.

• Unsupported

  We could not locate credible evidence to substantiate the claim at the time of publication.

• False

  Directly contradicted by primary sources and reliable documentation.

Verdicts reflect the cited evidence available at the time of publication. Where new credible evidence emerges, ratings may be revised. All revisions are logged transparently on the claim page. Readers may submit correction requests via our corrections page.

We focus on the claim, not on the person. We do not use language that implies unlawful conduct unless it is presented strictly as a sourced quotation with context. We avoid characterisations such as "lied", "fraud" or "corrupt" in our own editorial voice. Demonstration content shown across the site uses fictional sample names and is for illustrative purposes only.