WATCH THE SNARK: AI CAN NOW DETECT SARCASM ON SOCIAL MEDIA. Software developers refer to it euphemistically as “sentiment analysis.” Orwell had another term for it: “wrongthink.”
The University of Central Florida recently announced advances in creating Artificial Intelligence software that can detect sarcasm in social media and other online activities.
Artificial intelligence has seen rapid advances in logical data analysis and response, but it lagged in parsing emotional meanings and cues, especially within text. As Assistant Professor of Engineering Ivan Garibay explained in the press release:
“The presence of sarcasm in text is the main hindrance in the performance of sentiment analysis. Sarcasm isn’t always easy to identify in conversation, so you can imagine it’s pretty challenging for a computer program to do it and do it well. We developed an interpretable deep learning model using multi-head self-attention and gated recurrent units. The multi-head self-attention module aids in identifying crucial sarcastic cue-words from the input, and the recurrent units learn long-range dependencies between these cue-words to better classify the input text.”
Though the project was funded in part by a Defense Advanced Research Projects Agency (DARPA) grant, UCF claimed its purpose is to aid companies in marketing and selling products. Currently, many people are employed at larger companies to monitor and assess feedback and communications regarding products and services. The work is labor-intensive, and some that humans do best – until now, perhaps.
Nuance And Political Control
So, what can go wrong?
Quite a lot, given the record of recent abuse by government authorities in utilizing U.S. intelligence assets against citizens for political reasons. According to Vice News, the FBI has already been using software called “Babel X” that sifts through social media, as well as “dark web” and other sites, to glean sentiments of user profiles it compiles.
According to its makers, Babel X has powerful features of its own, including the most sophisticated sentiment analysis engine currently in production use:
“Babel X can also surveil millions of URLs including the deep web. The software can instantly translate over 200 languages, and can set up geo-fences around areas of special interest, and has highly customizable filtering options including for hashtags, emojis, handles, names, and keywords. Users can also filter for numerical sequences like credit card or social security numbers…
[Babel X] possess the most sophisticated sentiment analysis tool on the market. Derived from collaboration with top university linguistic programs, Babel Street boasts the ability to evaluate sentiment in 19 languages—far exceeding the capacity of any other competitor.”
Babel X was utilized to screen National Guard units that were deployed to Washington D.C. after the massive, mostly peaceful 6 January election fraud protest.
And in March 2021, at a “Future of Defense” symposium sponsored by The Hill, Senator Tammy Duckworth recommended that the Department of Defense employ tools like Babel X to monitor the social media habits of military members:
“It’s not a new thing, but I will tell you that I have seen over the last probably two decades this growing radicalization of a portion within the military. And I think part of it too comes with social media consumption.”
Given the DARPA funding of the UCF project and the documented government domestic use of Babel X, it appears likely that the latest advances in “sentiment analysis” will be leveraged to discriminate and abuse the political freedoms of American citizens.