This summer was extremely productive at MarvelousAI. As the presidential primary kicks back into full swing, we are laser-focused on tracking the daily information struggles around the candidates and their campaigns. So it seems a good time to take stock of our recent accomplishments.
In collaboration with the Credibility Coalition, we co-lead the development of the AMITT (Adversarial Misinformation and Influence Tactics and Techniques) framework for misinformation tracking and mitigation. AMITT is a stage-based framework inspired by MITRE’s ATT&CK framework. You can read about the work of the Misinfosec Working Group in our recent whitepaper.
Gender and Politics
In collaboration with Sarah Oates (UMD) and Lucina Di Meco (Wilson Center), we presented the study “Running While Female: Using AI to Track how Twitter Commentary Disadvantages Women in the 2020 U.S. Primaries” at the American Political Science Association.
We’ve opened up our candidate tracking system as an alpha release. You can now sign up to browse our unique combination of natural language analysis and publicly available publications, conversations, and media ratings as it matures.
Debate Narrative Tracking
We used our NLP tools to analyze the post-debate Twitter conversations, and worked with Aleszu Bajak (NWU / Storybench) on “Quantifying Twitter Attacks on Kamala Harris during and following the Democratic Debates.“
Public Speaking Engagements
We presented at the ReWork AI for Good Conference in San Francisco on Misinformation.
We were featured on the ReConsider Podcast to talk about Finding BS and Other Semantics Analysis in the 2020 Election.
It’s going to be a busy fall as the election heats up. We expect more conversation about the candidates, more attacks, and an increasing focus on the narratives most likely to suppress some voters and motivate others. At MarvelousAI we’ll be keeping a close eye on all of that. Get in touch if you’re interested!