logo

About

Back

Evaluating Media Bias

 
It's important to note that the evaluation of media bias is a very rough guestimate performed automatically from an AI Prompt. The purpose is not to provide a confirmed review of each individual source (there are over 800+) but rather, to enable the researcher to get some sense of potential bias based on this automated evaluation but each researcher should perform their own evaluation for anything that will be used in professional work or a published report.
 
For the purposes of transparency, the AI Prompt is provided below:
 
🤖 AI Prompt: Assess the source using its provided metadata and analyze it using criteria derived from Ad Fontes Media Bias Chart and Media Bias/Fact Check methodologies:
 
1. Factual Accuracy
Relies on primary sources/experts?{Yes/Partial/No]
Contains uncorrected errors? [None/Minor/Major]
2. Transparency
Clear authorship/editorial team? [Yes/Partial/No]
Discloses funding/ownership? [Full/Partial/None]
3. Sourcing Quality
Uses verifiable references? [90%+/50-89%/Below 50%]
Links to credible institutions? [Yes/Partial/No]
4. Bias Indicators
Neutral language vs loaded terms ratio: [3:1/2:1/1:1]
Acknowledges counter-arguments? [Yes/Partial/No]
5. Corrections History
Updates errors promptly? [Always/Sometimes/Never]
 
Scoring System (1-10)
9-10 = High Integrity
Rare errors, full transparency, expert sources
5-8 = Medium Integrity
Occasional issues, partial disclosures
1-4 = Low Integrity
Frequent errors, hidden funding, unreliable sources

Evaluating Source Integrity

 
It's important to note that the evaluation of source integrity is a very rough guestimate performed automatically from an AI Prompt.The purpose is not to provide a confirmed integrity review of each individual source (there are over 800+) but rather, to enable the researcher to get some sense of potential integrity based on this automated evaluation but each researcher should perform their own evaluation for anything that will be used in professional work or a published report.
 
For the purposes of transparency, the AI Prompt is provided below:
 
🤖 AI Prompt: Analyze the provided context to determine the integrity level of the source. Consider the credibility, reliability, and alignment with humanitarian and international development standards. Evaluate these aspects and provide a Score out of 5 for each qualifier:
 
Authority - Author/reputation indicators
Accuracy - Evidence of verification/primary sources
Bias - Neutral language vs loaded terms
Recency - Content freshness vs topic requirements
Transparency - Clear authorship/affiliations
 
Output format:
Provide a score out of 5 (where 1 equals low and 5 = high) for each of the five bullet points.
 

Evaluating H3 Potential

 
This is an automated workflow attempting to identify which sources are more likely to surface H3 Signals. Because it's automated and based on a UN Futures Impact-Probability Matrix, for some sources - this rating will be different than the "Horizon Rating" I might have applied initially when I input the source.
 
All things I need to review at some point - but for now, I've included both.
 
 
For the purposes of transparency, the AI Prompt is provided below:
 
🤖 AI Prompt: You are an AI tasked with analyzing sources using Horizon 3 criteria to identify emerging changes over a 10+ year horizon. Your expertise includes evaluating signal novelty, early detection potential, and disruptive impact profiles.
 
Follow these steps to perform the analysis:
1. First, assess the source's domain and tags for novelty by checking for non-traditional TLDs and niche topics.
 
  1. Next evaluate the content text for early detection potential, looking for speculative language and cross-disciplinary references.
     
  2. Finally, use the UN Global Pulse Impact-Probability Matrix to determine the disruptive impact profile.
 
Output format:
Provide a classification of the source's potential as High, Medium, or Low, followed by key indicators. Format the output as plain text without additional text, using the example:
" 🔵 High Potential | .bio domain | Quantum bioethics impact | 8mo first-mover advantage".