“In this report we present a system that can generate political speeches for a desired political party,” wrote Kassarnig in his recent publication. “Furthermore, the system allows to specify whether a speech should hold a supportive or opposing opinion.” The graduate student developed his impressive system by first creating a database containing some 4,000 political speech fragments from 53 Congressional floor debates. These segments were categorized by political affiliation and pro or con status, and ultimately, Kassarnig analyzed more than 50,000 sentences with an average of 23 words each.
This resulted in the AI’s ability to string together sequences of words, with a language model assuring grammatical correctness, and a topic model ensuring textual consistency. And while it’s not quite perfect yet, the following is an example of what the system has produced:
“Mr. Speaker, for years, honest but unfortunate consumers have had the ability to plead their case to come under bankruptcy protection and have their reasonable and valid debts discharged. The way the system is supposed to work, the bankruptcy court evaluates various factors including income, assets and debt to determine what debts can be paid and how consumers can get back on their feet. Stand up for growth and opportunity. Pass this legislation.”
Not too shabby! So while the AI may not replace our best speech writers or produce a State of the Union anytime soon, it certainly looks like it’s off to a good start.
