I used to believe when I was coaching that debate helped students in a number of areas that would be of lifelong value to them. There is public speaking, of course, something with which I had my own difficulties when I was younger. Sooner or later, in countless professions, one is called upon to make a presentation, and some professions are virtually endless presentations. Skill at this is definitely a skill for life, and debate is a great realm for learning it. And it is unlikely that anything about AI will replace that learning or that value.
Which leaves us everything else I thought students would be learning.
There's research, of course. The art of research has changed dramatically in my lifetime. When I was a debater back during the Van Buren administration, research was a process of manipulating physical media. We went to the library and used The Reader's Guide to Periodic Literature to find articles in magazines that we then dug out of the stacks. The information we gleaned from those articles we wrote onto 3*5 index cards, and those cards, in shoeboxes, comprised our research. In the first phase of the personal computer, which was a machine that was not connected to a Web that didn't exist yet, I acquired an encyclopedia on a CD, Microsoft Encarta, for my personal research. Then the internet and search engines came along, and the magic of Google's algorithm. And then there was Wikipedia, which got off to a bumpy and rather unreliable start, to put it mildly, before becoming quite useful. The unifying factor of all of these is that their manipulation required skill. That is, it could be said that, using the tools at hand, one was, or wasn't, a good researcher. And learning to be a good researcher was another area of lifelong value added to a debater's education.
Next up on the list of skills to be learnt was writing. To debate you needed cases, and if you wanted cases, you had to sit down and write them. In my experience, contemporary elementary school education does not come close to teaching students how to write. At best it teaches an approximation of what writing might be if one removed spelling, punctuation, grammar, and anything resembling lucidity. The writing of grammar school students is the equivalent of what we assume were the grunts and groans of the pre-Neanderthals: something was being communicated, but only at the most basic level. You can't communicate complex thoughts if you can't articulate them in the first place, and if you can't write them down, they remain unarticulated. Debating forced students, one way or another, to become decent writers, even if only of something that was going to be entirely spoken. (The exchange of written cases is relatively modern.)
And finally, there is the logic of argumentation. (There are other valuable skills in Speech and Debate, of course, but we're concentrating here on the ones most in present-day jeopardy.) There is the structure of claim/warrant/impact. There is the ability to listen to what someone is saying and to respond cogently. There is the process of weighing in one's own mind the importance of what is being said, and prioritizing accordingly. In other words, there is the art of debate itself. What makes a good debater is the ability to present their own position clearly, and to clearly refute the position of the opponent. You take what is said and turn it around or disprove it or show that it misses the point and that what you are saying outweighs it. Refutation is a combination of quick thinking and logic and very much knowing what you're talking about (which you learned from your research). And successful refutation is the key to debate success.
And here's the problem. When it comes to research, writing, and argumentation, AI at its simplest, most easily accessible form, is incomparably better at the first two than any high school debater, and a fair contender with the third one. (I'm not going to demonstrate any of this. Either you know it already, or if you don't believe it, you can try it for yourself.)
As for research, any AI agent can feed you a list of sources on literally any topic, saving you the need to poke around on Google or Wikipedia using only your wits and a galaxy-hitchhiking towel. If it is the first of the month and the topic has just been released, AI is probably not a bad starting point. It's the next progressive step in that research process I described earlier that used to begin with a physical trip to the public library. If this initial step in research is followed by careful analysis and evaluation of the results, I'm all for it. We would be using the presumably best possible research tool, combined with our own skill at evaluating research. Debate coaches can still envision educating a good researcher.
As for writing, I have fed an NSDA PF topic into an AI agent, and gotten back, in seconds, a perfectly acceptable set of both Pro and Con cases, complete with citations and timings to keep it within the speaking limits. Any novice could walk into a round with those cases and feel completely confident. Needless to say, with some juicing of one's queries, AI could produce some really excellent cases. Which means that, instead of learning to write cases, one would learn to use AI to write cases. This may or may not be a valuable skill—will anyone other than fiction authors actually write anything in the future?—but it certainly doesn't provide much education to a student in literal writing. It is, to writing, what asking AI to draw a picture is to art. It is not even close.
Finally, there's going into the round and debating. Here let us imagine the completely usual process of teams entering a round and immediately sharing their cases. (The art of flowing, I am told, is akin today to the art of etching Egyptian hieroglyphics.) What a team can do, as soon as they have the opposition case in hand, is feed that to their AI agent and have the AI agent provide logical rebuttals, complete with citations. This means that debaters don't have to consider what their opponent is saying and figure out responses: AI has done it for them. For that matter, they can ask their AI agent for some leading questions for Crossfire or CX. They can ask their agent to weigh the opposition arguments and tell them which are the strongest and which are the weakest and where is it that they should attack with the best chance of a favorable outcome. None of this communication with their agent will take them more than a few seconds.
And that is the state of AI today. (It will be getting better and better very quickly.) As for the research part, this is where the educator's skill comes into play. A debate coach in the 1960s had to teach a novice how to research. In 2026, a debate coach still needs to teach a novice how to research. The only thing that has changed is the actual "how." As for the writing of cases, again we rely on the educators. The debate coach can help students learn to write well, or the debate coach can skip that altogether and simply help students get their AI to write well. Those latter coaches will not be doing their students any favor...perhaps. But if the art of writing does go away, assigned to AI drones the way so many other things are being done by machines nowadays, well, I don't know. E. B. White's words may be prophetic: "It is not often that someone comes along who is a true friend and a good writer." (That line still brings a tear to my eye.)
It is the use of AI in actual rounds, as an agent assisting debaters in their arguments, with which I take issue. Allowing use of AI in rounds means that it is not inconceivable that two teams will be facing off arguing not their own ideas and responses but the ideas and responses fed to them by their AI. This is not fantasy: it is absolutely possible. Worse still, one side could be relying on their own wits while the other side is arguing with an unseen assistant at their side, an assistant with the full access to the might and main of the internet, accessible in microseconds. If both teams use AI, what's the point of having the humans? If one team uses AI, how is that even remotely fair?
I will be petitioning my local league to ban AI in rounds. This includes, by the way, speech. I have tested this: a basic AI agent can create an absolutely perfect extemp speech in a handful of seconds, complete with timings and citations and speaking hints. Which means that an extemper who once spent precious minutes researching followed by precious minutes preparing a speech followed by precious minutes memorizing it, can now simply spend all those precious minutes on the memorization of something they had virtually no hand in creating. Where's the value in that? Where is the competitive fairness? Are we ranking speakers or are we ranking AI programs?
To be honest, I think that AI is the greatest challenge that coaches face today. AI needs to be understood and used wisely by coaches. And it does need to be used: It is real, it is not going away, and there are ways of using it that will provide good education to students. But there are also ways of using it that are, in a competitive environment, cheating. Let me emphasize that. Use of AI in rounds, which is the accessing of outside agents, is as much cheating as having your teammates in the back of the room provide you responses during the round in your Google doc. The only difference is the inherent humanity. Cheating is cheating no matter how you slice it.
If I successfully pass a rule prohibiting AI in rounds, can it be enforceable? Probably not. I can only offer the honor system as enforcement. Given that I'm proposing this rule in our local Catholic Forensic League, I don't think it is too much to ask that we rely on that honor system. The students who are going to cheat are always going to find a way to do so—it's not as if cheating is unknown in our activity—but the vast majority of students, teams and coaches and judges, will be good citizens, adhering to the rule, and will gain the reward of achieving our speech and debate educational goal. And if we don't provide some guardrails against misuse of AI in our activity? First off, we can just dump extemp as totally worthless as an event, and before long, as students learn to train their AI (I imagine some eager entrepreneur coming up with Chat1AR before too long, available for just $9.99 a month plus, well, your immortal soul), we can dump PF, LD, CX and BQ. (I know little of how student congress works, but I would imagine it too is in jeopardy.)
Do any of us want to be around when Claude wins the TOC? In all events?
Addendum: NSDA does ban AI in competition. NCFL does not. So this does, obviously, apply to CFL tournaments specifically. Which is why it's a rule I'm presenting to my local CFL league. I was rather shocked to find that use of AI in competition was acceptable when I worked at their MS virtual tournament. I wrote this rather long piece, which would all be relatively obvious to probably the vast majority in the activity, in the hopes of getting their national organization on board with this.