KI-MentorMedia Rating Platform
News can feel convincing before anyone has checked its quality. KI-Mentor makes trust signals visible: source ratings, manipulation scores, and recent analysis history — for anyone who needs to judge media quality faster and with evidence.
KI-Mentor rates news outlets by how densely their articles use persuasion techniques. 131 outlets are tracked from 435 analysed articles; 6 have passed the 10-analysis threshold and carry a published A+ to F grade. A high score means few techniques were detected. Currently br.de leads with B- (53.3/100) and zeit.de ranks lowest at C- (37.6/100).
AI-generated analysis. Rating and reasoning are produced automatically and are not individually reviewed by an editor.
| # | Outlet | Rating | Vol. | |||
|---|---|---|---|---|---|---|
| 1 | B- | 13 | ||||
| 2 | C+ | 79 | ||||
37ManipulationAnkunft zahlreicher Migranten: Spanien entsendet Soldaten in Exklave Ceuta | tagesschau.de | ||||||
| 3 | C+ | 13 | ||||
| 4 | C+ | 21 | ||||
63ManipulationCeuta gestürmt: Kommen die Migranten jetzt nach Deutschland? | Politik | BILD.de | ||||||
| 5 | C | 15 | ||||
| 6 | C- | 10 | ||||
68ManipulationRobert Koch-Institut: RKI meldet fast 10.000 Hitzetote seit Jahresbeginn | DIE ZEIT | ||||||
Analysed, not yet rated
69 outlets · under 10 analysesThese outlets have been analysed, but need more analyses before a dependable rating is possible — they are rated from 10 analyses onward. The values so far are visible, but they are not a rating yet. Closest first.
| Outlet | Vol. | |||||
|---|---|---|---|---|---|---|
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
45Manipulation
SCHUNK – Technologiepionier in Spanntechnik, Greiftechnik und Automatisierungstechnik. | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
41ManipulationWÜRTH » Ihr Spezialist für Handwerk und Industrie | WÜRTH | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
| 1/101 of 10 analyses needed for a rating | ||||||
How Does the Rating Work?
Every analyzed article from a media outlet contributes to its overall rating. The platform automatically evaluates texts for manipulation techniques such as emotional language, logical fallacies, statistical distortions, and rhetorical patterns. The result is a transparent, comparable quality index.
Real-Time Analysis
6-step analysis detects over 30 manipulation techniques — from emotional language to subtle framing patterns.
Scientific Scoring
Scoring based on density, intensity, coverage, and compound effects of techniques — comparable to financial rating methodology.
Independent & Transparent
No editorial influence. Every rating is traceable — techniques, text excerpts, and scores are fully disclosed.
Rating Scale Methodology v2.3
Markedly fewer techniques than average
Below to near average
Average to above-average technique density
Markedly above average
Very high technique density
Not Rated
Frequently Asked Questions
What is KI-Mentor?
KI-Mentor is an independent platform that rates media outlets using automated text analysis. Much as rating agencies assess companies, KI-Mentor makes the linguistic quality of news outlets comparable.
How are the ratings calculated?
Every analysed article is checked for more than 30 persuasion techniques. The score rests on four factors: density, intensity, variety and how techniques compound. A fixed, versioned calibration makes scores comparable across text type, length and analysis depth — 0 means no techniques were detected, 50 is roughly a typical text. The fewer techniques detected, the better the rating.
When does an outlet get a grade?
A grade from A+ to F is published only from 10 analysed articles onward. Below that the outlet appears in the list but without a rank and without a grade — a single analysis lands at an extreme and would not be a dependable statement about a whole outlet.
Which media outlets are rated?
The current focus is on German-language media outlets. Ratings update continuously as new analyses arrive — any article URL can be submitted for analysis.
Is the rating neutral and independent?
Yes. Analysis is fully automated with no editorial influence, and there is no paid placement: an outlet cannot buy, change or remove its grade or its position. The methodology is public.