(Molly Bruns, Headline USA) A recent announcement by the Office of the Director of National Intelligence revealed that the Intelligence Advanced Projects Activity is developing a program to expose anonymous writers.
IARPA will be using AI to analyze anonymous writers’ style, and claims that a writer’s style “is seen as potentially being as unique as a fingerprint.”
“Humans and machines produce vast amounts of text content every day. Text contains linguistic features that can reveal author identity,” IARPA stated.
According to Big League Politics, if the group succeeds it believs that the program could identify a writer’s style from multiple samples and adjust the patterns to increase people’s anonymity in writing.
“We have a strong chance of meeting our goals, delivering much-needed capabilities to the Intelligence Community, and substantially expanding our understanding of variation in human language using the latest advances in computational linguistics and deep learning,” declared the Human Interpretable Attribution or Text Using Underlying Structure (HIATUS) program manager Dr. Timothy McKinnon.
IARPA also said the HIATUS program would not only create explainability standards for the program’s AIs, but will also be capable of fighting foreign influence activities, defending writers whose work may endanger them and help identify counterintelligence.
The program will be able to identify machine generated versus human being written text.
Cindy Harper of Reclaim the Net said it’s possible that a writer’s style could be “as unique as a fingerprint.”
“It is not IARPA’s work to turn HIATUS into something usable,” Harper explained. “The agency’s work is only to develop the technology.”
It is not known when IARPA, a government organization, will put this project into effect, or if it will be used for strictly national security and intelligence purposes.
No matter when, it seems as if this technology could also be used to discover the identities of controversial writers who publish under pen names.