ChatGPT, is This Real?

I’d just opened my computer after drawing a question in the IX outrounds of the UT Longhorn Classic when I noticed, completely by accident, the screen of a fellow extemper who was set to compete as one of the speakers in my room. Although I only spared it a brief glance, it showed everything from fiery rhetoric to completely fleshed-out points from our one and only ChatGPT. 

They broke. I, unfortunately, did not.

Granted, it’s unlikely that their usage of artificial intelligence was the sole reason for their success, but the implications of using it in the most popular limited-prep event were concerning, to say the least. Other articles featured in the EIF blog have already speculated the possible uses of AI in speech and debate; now, it’s time to talk about how it’s destroying the true purpose behind these events in real time.

I was privileged enough to be introduced to extemporaneous speaking much earlier than many of my fellow competitors, as our district supported a middle school speech and debate circuit that emulated many of the events found in high school. There, I was taught efficient prep strategies, performative drills, and myriad other solutions to improving in the event, and extemp became a way for me, along with thousands of other people, to challenge their creativity, intelligence, and skills. Answering questions on everything from Japan’s upcoming elections to Brazil’s climate change efforts — or lack thereof — not only helped high school students across the nation develop critical thinking skills, but also prepared a future generation of Americans by getting them well-versed in both domestic and international politics, economics, and social issues.

But ever since ChatGPT was introduced in late 2022, some of the most fundamental purposes and characteristics of extemp are now in jeopardy. Despite the occasional news story of how one language-based AI or another had let it slip that their ultimate goal was to rule the world, several extempers have begun using ChatGPT, Bard, and several other models to transform how they utilize their prep time. For example, while most students come up with a unique umbrella answer and three supporting points within the first five minutes of their prep, AI can cut the time it takes to do that to a mere 30 seconds. Or, if competitors really want to hit the audience with passionate rhetoric, AI can provide over 15 ways to do so in a couple seconds alone. Cutting corners in preparing an extemp speech not only allows for more comparative practice time in delivering it during prep, but also directly insults the other extempers in the room and across the country. While some may argue that, as AI is available to all, it should be fair game as “part of the Internet,” this unequivocally removes the innate creativity and skill measurements used to gauge the merits of a competitor, which not only results in unfair breaks, but also threatens the very purpose of extemp as a whole: to help future leaders grow. This detriments extemp on multiple levels: a competitive one, a social one, and most importantly, a moral one.

Unfortunately, for those who agreed with the people that asked me “Why didn’t you just report [that competitor]?”, they’re in for a rude awakening. Not only do several state rulebooks, such as the Texas Forensic Association’s (TFA’s) Constitution, currently include no mention of the use of generative AI, but the 2023-2024 NSDA Manual actually allows artificial intelligence like ChatGPT to be used in “guid[ing] students to articles, ideas, and sources,” so long as the original source stating the evidence used in a speech is available upon request. Reading between the lines, this allows competitors to take points and rhetoric provided by generative AI and cite them as something completely original; after all, no one would think to question the originality of impactful phrases and entire points. This, quite obviously, leads to a very real and very near future in which more and more competitors take ideas provided by artificial intelligence and cite them as their own, leading to an event in which even national finalists may, at base, only be squawking parrots that regurgitate the words of an algorithm. Many of us regard the idea of AI taking over the world as nothing but cynical science fiction. But as a speech and debate community, we fail to realize that it’s already happening under our very noses.

SOLUTIONS

How can we prevent this? Most extempers, those who still use the supercomputer in their skulls rather than on their screens, pride themselves on their ability to find answers to daunting problems. This should be no different.

Starting from the ground up, extempers across America need to restore the integrity extemporaneous speaking once had. As a competitor, limit yourself to original ideas, original sources, and original rhetoric. To be frank, if a competitor needs literal AI to tell them what to say, what does that show about their true ability and dedication? If a competitor relies on content generated by a language model that, at the time this article was published, admitted it was only up to date on current events that occurred in January 2022, that already speaks volumes about their confidence in themselves. No one in extemp needs the help of artificial intelligence to get them the 1. Bring back the authenticity of extemporaneous speaking by relying on nothing but the sources you find, the ideas you come up with, and the eloquence of your own speaking style.

Even with the efforts of the majority, the unethical actions of some still necessitate much larger actions taken not by competitors, but by coaches who have the power to write the rules. To coaches around the nation, we as competitors already appreciate and love you for all of the effort and faith you put in the program. We’re just asking for a bit more. Schools far and wide have already banned the use of AI in homework assignments, major essays, and more. It’s time that competitions did the same. My fellow teammate and competitor Sasha Morel wrote of a proposed rule change in the NSDA Unified Manual that would transform extemp in roughly the same way that AI is already starting to do. If that rule change, which has outraged extempers like myself across the country, can be seriously considered by NSDA coaches, then a rule that outright bans the use of artificial intelligence in any facet of extemporaneous speaking should be seriously considered as well. I have absolute faith that proper enforcement mechanisms, similar to source checks and draw moderators, are completely possible. As the speech and debate community watches artificial intelligence branch into more and more aspects of life, let’s preserve the best parts of extemp — the originality, the personality,and the integrity of speakers — by keeping it out of our competitions.