AIUNT (Latin: What do they say?) listens to popular opinion about everyday topics by scientifically examining online comments from selected news outlets.

It sounds simple, but it can be a powerful tool for understanding: The words that people and publications use reveal larger trends and the formation of public opinion that have profound and far-reaching effects on behavior, politics, and policy.

Aiunt empowers users to analyze public attitudes by harnessing the power of machine learning. It continuously harvests articles and comments from selected online sources, preparing balanced and relevant samples, detecting potential biases, extracting hidden correlations, and displaying immediate results in visual form with the full range of statistical significance.

  • R
  • Weka
  • TM
  • OpenNLP
  • D3.js

The users set its search parameters, and Aiunt does the hard work in the background. It provides intuitive interfaces, so users can make rich analytics with no coding required, including sentiment scores, intuitive graphs, and frequency data. In addition to analyzing online news content, Aiunt can aggregate social media discussions on particular issues to establish empirical evidence and enable easy generation of recommendations and predictions.

Jefferson Institute created Aiunt in 2016, with financial support from the Knight Foundation. In its initial case study monitoring election-related content, Aiunt followed and analyzed the online discussions at a wide range of media outlets: Politico, Breitbart, Daily Kos, The Washington Post, Hot Air, Al Jazeera, Fox News, Info Wars, KOLO-TV, New Hampshire Public Radio, Red State, The Blaze, The Huffington Post Blog, and The Iowa Republican. By analyzing one or more sources at the same time, Aiunt was the best address to start research on what people were saying about a hot topic during the election period.