Source Trust Database
Source Trust Database
Public accountability for every news outlet in the system. Trust scores are computed from article integrity, framing pressure, and correction rates. No outlet gets to hide behind their brand.
Outlets Tracked24
High Trust6
Low Trust6
Articles Analyzed3,993
Trust 75+
Trust 55-74
Trust 35-54
Trust <35
| Source Outlet | Trust | Reliability | Articles | Integrity | Framing | Corrections | Spin Pattern |
|---|---|---|---|---|---|---|---|
| NPR News | 74 | HIGH | 101 | 70 | 37 | 0 | Moderate framing |
| TechCrunch AI | 74 | HIGH | 36 | 70 | 38 | 0 | Moderate framing |
| Daily Caller | 71 |
How Trust Scores Work
Integrity (50% weight)
Average evidence-integrity score across all articles from this source. Measured at ingestion time by the AI bias analyst jury. Higher = more claims backed by verifiable evidence.
Framing Pressure (30% weight)
How much loaded language, euphemisms, and narrative framing the source uses. Lower framing = more factual reporting. 100 - biasScore contributes to trust.
Correction Rate (20% weight)
Percentage of articles from this source that have been corrected. More corrections = lower trust. Sources that silently edit without logging corrections are flagged.
Disclaimer: Trust scores are algorithmically computed from AI-assisted analysis of source material. They are signals, not verdicts. Always read the evidence and decide for yourself. A source with a low score may still publish accurate reporting — and a source with a high score may still spin.