HIATUS- Human Interpretable Attribution of Text using Underlying Structure

  • Leader
    October 22, 2022 2:50 PM PDT

    Intelligence Summary

    The HIATUS program aims to develop novel human-useable systems for attributing authorship and protecting author privacy. Authorship attribution capabilities address many Intelligence Community (IC) needs, including combating sophisticated malicious information campaigns online and identifying counterintelligence risks. Authorship privacy capabilities protect authors whose writing, if attributed, could place them in danger.

     

    The "right" says this is AI that can expose anonymous writers. In other words, apparently the intent of this AI is to eliminate your AKA.

    Summary

    Humans and machines produce vast amounts of text content every day. Text contains linguistic features that can reveal author identity. To support and protect the IC mission, the HIATUS program’s objective is to develop multi-language-capable tools to attribute authorship and protect author privacy. These tools must implement novel explainable Artificial Intelligence techniques to provide trustworthy and verifiable results to human users regardless of author background or document genre, topic, and length.

     

    IARPA says that an author's identity will be protected. Interesting technology. What do you think?

    The Intelligence Advanced Research Projects Activity invests in high-risk, high-payoff research programs to tackle some of the most difficult challenges of the agencies and disciplines in the Intelligence Community (IC).

    This is who IARPA is:
    IARPA collaborates across the IC to ensure that our research addresses relevant future needs. This cross-community focus ensures our ability to address cross-agency challenges, leveraging both operational and research and development expertise from across the IC, and coordinating transition strategies with our agency partners.