Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Measuring and Fostering Peace through Machine Learning and Artificial Intelligence

P. Gilda, P. Dungarwal, A. Thongkham, E. T. Ajayi, S. Choudhary, T. M. Terol, C. Lam, J. P. Araujo, M. McFadyen-Mungalln, L. S. Liebovitch, P. T. Coleman, H. West, K. Sieck, S. Carter

TL;DR

The paper addresses measuring peace in public discourse and promoting peaceful media consumption using ML and AI. It first develops news-text peace classifiers using embedding-based representations and multiple neural architectures, then pivots to emotion- and context-aware analysis (GoEmotions and LLMs) for YouTube transcripts due to cross-domain gaps. A real-time Chrome extension, MirrorMirror, is developed to give users feedback on media tone, with human-expert benchmarks guiding evaluation and a plan to open-source the tool for creators, researchers, and platforms. Results show strong cross-dataset performance for written news, while YouTube content requires contextual, multi-stage AI reasoning, with LLMs achieving higher agreement with human judgments, suggesting practical potential for reducing polarization in online discourse.

Abstract

We used machine learning and artificial intelligence: 1) to measure levels of peace in countries from news and social media and 2) to develop on-line tools that promote peace by helping users better understand their own media diet. For news media, we used neural networks to measure levels of peace from text embeddings of on-line news sources. The model, trained on one news media dataset also showed high accuracy when used to analyze a different news dataset. For social media, such as YouTube, we developed other models to measure levels of social dimensions important in peace using word level (GoEmotions) and context level (Large Language Model) methods. To promote peace, we note that 71% of people 20-40 years old daily view most of their news through short videos on social media. Content creators of these videos are biased towards creating videos with emotional activation, making you angry to engage you, to increase clicks. We developed and tested a Chrome extension, MirrorMirror, which provides real-time feedback to YouTube viewers about the peacefulness of the media they are watching. Our long term goal is for MirrorMirror to evolve into an open-source tool for content creators, journalists, researchers, platforms, and individual users to better understand the tone of their media creation and consumption and its effects on viewers. Moving beyond simple engagement metrics, we hope to encourage more respectful, nuanced, and informative communication.

Measuring and Fostering Peace through Machine Learning and Artificial Intelligence

TL;DR

The paper addresses measuring peace in public discourse and promoting peaceful media consumption using ML and AI. It first develops news-text peace classifiers using embedding-based representations and multiple neural architectures, then pivots to emotion- and context-aware analysis (GoEmotions and LLMs) for YouTube transcripts due to cross-domain gaps. A real-time Chrome extension, MirrorMirror, is developed to give users feedback on media tone, with human-expert benchmarks guiding evaluation and a plan to open-source the tool for creators, researchers, and platforms. Results show strong cross-dataset performance for written news, while YouTube content requires contextual, multi-stage AI reasoning, with LLMs achieving higher agreement with human judgments, suggesting practical potential for reducing polarization in online discourse.

Abstract

We used machine learning and artificial intelligence: 1) to measure levels of peace in countries from news and social media and 2) to develop on-line tools that promote peace by helping users better understand their own media diet. For news media, we used neural networks to measure levels of peace from text embeddings of on-line news sources. The model, trained on one news media dataset also showed high accuracy when used to analyze a different news dataset. For social media, such as YouTube, we developed other models to measure levels of social dimensions important in peace using word level (GoEmotions) and context level (Large Language Model) methods. To promote peace, we note that 71% of people 20-40 years old daily view most of their news through short videos on social media. Content creators of these videos are biased towards creating videos with emotional activation, making you angry to engage you, to increase clicks. We developed and tested a Chrome extension, MirrorMirror, which provides real-time feedback to YouTube viewers about the peacefulness of the media they are watching. Our long term goal is for MirrorMirror to evolve into an open-source tool for content creators, journalists, researchers, platforms, and individual users to better understand the tone of their media creation and consumption and its effects on viewers. Moving beyond simple engagement metrics, we hope to encourage more respectful, nuanced, and informative communication.
Paper Structure (24 sections, 4 figures, 2 tables)

This paper contains 24 sections, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methods: News Data
  3. Strategy
  4. Neural Network Classification of Peace
  5. Data Preparation
  6. Model Architectures
  7. Training Procedure
  8. Results: News Data
  9. Evaluation of Neural Networks for Peace Classification
  10. Model Performance on Test Data
  11. Cross-Dataset Generalization
  12. Comparison Between Network Architectures
  13. Cross-Domain Generalization to YouTube Transcripts
  14. Methods: YouTube Data
  15. Strategy
  16. ...and 9 more sections

Figures (4)

  • Figure 1: Neural network training and evaluation pipeline for peace classification.
  • Figure 2: Pearson correlation coefficient $r$ between the human coders and predictions of the AI models on the dimension: compassion---contempt. We used the models: Gemini 3 Pro Preview, Gemini 2.5 Flash, GPT-5.1, GPT-4o, RoBERTa, and GoEmotions. +R indicates hybrid models with RoBERTa.
  • Figure 3: Pearson correlation coefficient $r$ between the human coders and predictions of the AI models for all 5 social science dimensions.
  • Figure 4: Prototype of the MirrorMirror user interface.