In the 1990s the Odyssey was rebranded for the silver screen and titled Terminator 2: Judgement. Odysseus was named Sarah Connor (Linda Hamilton),  and her long  journey home  is now aided by her tough guy teenaged son and a previously evil cybernetic organism from the future (played by Arnold Schwarzenegger) now sent to protect her and her son. The once villain now hero is reprogrammed for good and sent back in time to protect the very humans he was once created to destroy.  This academy award winning film (for visual effects,sound mixing and makeup BUT not for best movie of all time about time travel and robots) remains in the cultural zeitgeist when talking about technology taking over. In summary, Skynet (the fictional defense company in the Terminator saga) represents the ultimate man vs machine fear; the machines see the biggest threat to peace is humankind since all wars are started and fought by humans, therefore peace can only be achieved when all humans are destroyed.  Scary!

As a teaching fellow, I can’t help but think about terminators when I hear AI writing software names like Chat GPT. Is this new and therefore scary at first or is it a dangerous self-aware monster that’s going to destroy the world of writing as we know it? In AI for Editing by Nupoor Ranade, the author allies with AI services and invites students to partner and explore AI software. The use of AI as a writing enhancement tool reveals the strengths and weaknesses of the software. In terms of summarization most platforms were adequate but problematic in accuracy, misrepresenting the information. When it came to analysis, AI was able to set up the frame of an argument and reference the topic described, but no real supporting evidence to back-up the argument. A common problem within AI writing software is hallucinations or made up sources/citations by AI to persuade the reader.  This a dangerous trick of technology that can go unnoticed by uninformed user.

Ethical Considerations in Courtroom Drama

According to this story from Forbes.com, Steven A. Schwartz, a seasoned lawyer (30+ years of courtroom experience) used Chat GPT to create arguments for his client’s case. Schwartz asked the AI software about the cases/sources it named, and the software reassured Schwartz that the information was real. It was in fact, not real. The program hallucinated, fabricating cases that never happened, to achieve the objective of justifying arguments. Now this lawyer faces severe consequences, claiming he didn’t know the software would be capable of doing something so dishonest. What’s frightening about this is that it was caught, posing the idea that there could be lawyers doing this same thing, only going unnoticed. The fabric of the legal system is only one of many areas that AI writing could appear to partner with, but unknowingly betray the user through their own ignorance.

 The literacy of AI expands as the technology outpaces the learning by teachers and administrations to navigate this new real with caution and prevent this tool from becoming the destructive terminator-like machine that eliminates the humanity from writing. What can we do as teachers of rhetoric? We can treat this new tool with caution as we explore it with our students, making them aware of these flaws within the system. Making our students AI literate will hopefully enhance the usage of this tool to produce better writing. Getting students to step away from the software and develop their own writing muscles is a choice they must make and endure. One major challenge is not failure, that’s a natural part of life, it’s dealing with failure and adjusting for success. One major tool I use all the time is the “terrible example.” Where I show an awful resume, infographic, or poorly constructed website. I have the students fix it as a group, in a discussion (that ultimately turns into a feeding frenzy) where we constructively repair these terrible documents using the concepts we are about to define or have been exposed to. Students start to notice the typography changes, the poor spelling, the horrible document layout colors and their reactions help create a lively discussion about repairing documents, to fit the mold of what something abstract can fulfill.

AI writing software is perfectly fine as an editor on a rough draft, formatting citations, and on cleaning up that old resume. However, AI writing doing the heavy lifting of constructing the critical thought and supplying evidence (sometimes false) creates a dangerous reality that like the terminators warned us, removes humanity to prevent human mistakes. We must reprogram these bots, in order to ally with them, and fight the good fight, the fight of keeping humanity alive within the written word.

In Dante’s epic poem The Inferno (an epic poem that everyone should pretend to read), after becoming lost, he follows poet Virgil only to be led into the nine circles of Hell. There’s a whole to-do that follows about the soul, that I will not spoil. Dante’s poem inspired films, books, songs, and as of 2009, the good people at PlayStation gave us the long-awaited video game. Speaking for every 14^th century reader when I say, “It’s about time.” The fans or Infernheads are finally rewarded some 8 centuries later (see trailer below):

https://www.youtube.com/watch?v=ysCrkeoKdnY

Dante’s Inferno Video Game Trailer

Today I cannot help but think about Dante’s epic journey into darkness, as I begin to explore Artificial Intelligence (AI) generating large language models (LLM) of text like Chat GPT. First and foremost, I don’t want to paint myself as a Luddite having a tantrum about the inevitable changes of life and technology. It never ages well in writing. See Henry David Thoreau lose his mind over trains going 35 mph. Thoreau more eloquently said, “Not until we are lost, do we begin to find ourselves.” And I can fully admit that although I am lost within the realms of AI writing, I am looking forward to finding myself through this explorative quest.

 Barely capable of anything myself, I can always ask Siri, or her cousin Google for help, and they withhold judgement and help me in their haunted and toneless voices.

My pro-technology stance remains tied to comfort of not needing to know everything. However, I find myself concerned. Frankly I’m more than a little concerned about how fast AI is advancing and how little we know as educators. As we begin to take this journey, I offer classroom discussions and reflections based on two ideas

1. AI’s lack of representation
2. AI’s production of canned/poor quality writing

Representation

In An Introduction to Teaching with Text Generation Technologies by Tim Laquintano, Carly Schnitzler, and Annette Vee, the authors examine the state of AI writing, dating back to the 17^th century. The authors reveal that AI technology can exclude many historically marginalized groups, simply because the software looks toward what is available, and sadly the published work is not diverse. As representation is missing within technical communication, this exclusionary element of AI writing technology poses the threat of continuing the divide. I brought this sad fact to the attention of my class in a 10-minute writing prompt, followed by discussion in my current technical communication classroom.

Prompt:

 If AI writing is only representing groups that are already prominently represented, how do we correct this?

Students concluded that the software itself is still in its infancy, but it does learn through exposure to content. If the AI world is exposed to more voices, writers, perspectives then the software itself will learn and represent them accordingly. Many students also voiced a desire to enter the field and personally correct this error of under-representation.

 Quality

Using AI as an educative tool to guide the expression of ideas, instructors can further the writing process into a new realm of creativity and take young minds forward into the craft of words. This noble and pious use of the machine another resource can help students beyond the classroom. That same tool can be manipulated to fulfill the sins of those given to temptation to cheat, circumnavigate learning and end ultimately end up in a purgatory for plagiarists. Within this is something that comes with the territory, some people will beat the system. Those people will not learn, and you can spend time trying to decode auto-essays, but this is a rabbit hole I prefer not to fall into. The easiest way to spot an AI writer is to remove technology, and if a student is possibly using AI for evil, hand them a blank paper and a pen and tell them to handwrite it in front of you. This is something I have told my current Technical Communications sophomore class from day one. After repeatedly seeing their smiling eyes downward affixed to their screens, I decided to make them write the old-fashioned way. Once they’ve crafted outlines of ideas and getting anything onto the page THEN we can begin to play around with prompts in Chat GPT to expand our initial ideas.

In Douglas Eyman’s Text Generators in Technical Communication: Summarizing Technical Documents students are prompted to explore the landscape of AI writing software. Using various AI writing platforms, graduate level students are exploring the limits, depths, and results of prompting these programs to generate accurate and effective technical summaries. The limitations of these platforms for creating quality work became evident and showed students that not all AI is created equally. Eyman’s student found, “Some systems used the main document headings to produce a summary, while others drew from text provided in later paragraphs or each section; because we were familiar with the material, we could quickly see which variation provided a more accurate summary.”

Using this idea to emphasize quality and recognition of accuracy, my students looked at an auto-generated cover letter template and described it as an empty template written without feeling. The students were able to spot the hollow nature of the writing quickly, basic sentence patterns and ultimately concluded that the AI cover letter would not stand out in a pile of applicants. We discussed how this template if filled in, would also fail in the job market because it had too plain a nature.  As a class we added possibilities for the writer to explore sections with the “Topic, Claim and Support” format. Then I brought the one word that Dante would be proud of me for, “Soul.” Quality writing has soul, humanity, personality, and authenticity. It cannot simply be filled in like Mad-Libs and given to an anonymous HR rep. By injecting self into the writing, students breathe life into the words, reflect their tone and creates a human-to-human connection that gives you a chance to compete for the job.