News Bias Detector
An LLM-based AI Service - Available on SingularityNET
News articles often suffer from bias and manipulation, making it hard to find objective reporting. This service is designed to detect bias in news articles and ensure that you can rely on more credible information.
Service Description:
Each submitted news article is analyzed and given a bias score from 0 (extremely biased) to 1 (non-biased).
Our 2nd version is currently online, which we have converted to the LLM from OpenAI. The advantage of LLM over NLP models in this case is that subliminal biases in particular are better recognised. With this update, in addition to the bias score, there is also an explanation of how the score is determined.
The following reference scores can be used to interpret the output values:
0.0 - 0.6 = highly biased
0.7 - 0.8 = moderately biased
0.9 - 1.0 = not biased
How to use it:
This service is published and available on SingularityNET's AI marketplace. To use this service, you need to create an account and connect your MetaMask Wallet with the portal.
Everyone has the opportunity to request 3 free demo analyses. After that they only cost 0.1 AGIX per call (currently around 0.05 USD)
Link to the services:https://beta.singularitynet.io/servicedetails/org/fafe2dc448cd4e0bb581d2c29020282f/service/trustlevel-bias-service
Github Documentation:
If you are interested in using this services directly via our API, please contact us via info@trustlevel.io
Last updated
