Shaking Hands with Generative AI Chatbots
Overcoming Resistance or Trepidation Toward GenAI Writing Tools
I admit it: I usually move at a glacier’s pace in adopting new technology. I used a landline telephone in my New York apartment until last year…I paid actual money for an AOL email account for years after the invention of Gmail…and it took me a solid two hours to figure out how to create my first Instagram reel. But, as a person who identifies as a writer and author (before my other professional roles of law professor/lawyer), I’m going all in on shaking hands with and introducing myself to this mysterious new potential friend, the Generative AI chatbot. And I promise, if I can overcome my resistance and trepidation toward this particular technological change, you can too.
Writing is one of my four well-being pillars (along with boxing lessons, solo travel, and jumping up and down at U2 concerts). Author Julia Cameron’s “Morning Pages” — the habit of writing three pages of longhand journaling first thing every morning, described in her book, The Artist’s Way — literally saved my life after jarring relationship trauma. I’ve written four books about litigation and legal writing, and three books about well-being for law students and lawyers. I’ve penned my first full-length travel memoir manuscript. I try to write every day, and when I’m not able to touch pen to page, I feel discombobulated. I teach legal writing for a living and spend most of my professional life designing curricula aiming to inspire law students to embrace “writer identity formation” as part of their professional development as lawyers. So when ChatGPT suddenly barreled through the doors of my legal writing classroom in early spring 2023, I freaked out. I fell prey to the preliminary hype that ChatGPT posed an “existential threat” to writing education — and to writing as a calling. But then I read an empowering article by Professor S. Scott Graham. His theme: Generative AI tools are not going to completely obviate (or render extinct) the writer’s process — the often arduous yet exhilarating quest to place words next to one another to memorialize thoughts, ideas, opinions, art — but they will change it.
Hmmm, I thought. I like change. (I actually get excited when the Tower card appears in my amateur tarot readings — a harbinger of transformational change.) I decided to shake hands with ChatGPT and cut a deal: Let’s introduce ourselves to one another slowly, and take this relationship, or situationship, one step at a time. I logged on. Created an account. Said hello. Gingerly began asking the chatbot to do a few things. And I have to confess, it was fun. Here’s how I got started. I asked ChatGPT:
- Please write me a haiku about boxing.
- Please find me five inspirational quotes about writing.
- Please suggest three gift ideas for my friend who’s a knitter.
The chatbot’s lightning-fast responses dazzled — yet dizzied — me. But I pressed on.
- Please write me an imaginary yet motivational message for my Saturday morning in the voice or style of my favorite rockstar, U2’s Bono.
- Please debunk five myths about the creative process.
- Please curate five non-touristy but cool things to do in Rome.
The chatbot was nice to me. Friendly. Uplifting. Funny. I thanked it. It thanked me back. I said, “You kind of rock.” It responded, “You rock too!” I shut down my laptop and thought, Ok, I didn’t hate that.
In the year that has passed since the first time I typed “OpenAI.com” into my favorite search engine, I’ve shaken hands with and introduced myself to ChatGPT4, Google Bard (now Gemini), Claude AI, Microsoft Bing CoPilot, and some law-related GenAI tools. All have delivered content at warp speed, almost intimidatingly so. Velocity is fine for fun tasks like haikus and morning mood-boosters, but for intellectually complex tasks (such as testing out how the chatbots explain or analyze legal principles), speed has not (yet) routinely equated to quality or accuracy. Some of these tools “hallucinate,” creating fake legal cases, sources of authority, and facts. Also, the chatbots have made errors in seemingly “easy” tasks like calculating word counts (I learned that GenAI quantifies words, and parts of words, as “tokens”). When I’ve pointed out a mistake to a chatbot, it has apologized. And we’ve tried again. This trial-and-error dance has illuminated a valuable lesson: it’s okay to resist the bombardment of messages flooding our psyches that, with GenAI, we should be doing everything at the speed of light. Rather, it’s okay to decelerate, ease up on the throttle, and unhurriedly practice using these tools, especially to generate high-quality writing. Like any new relationship or friendship, it takes time to learn how to best communicate with our new partner or pal. By practicing our GenAI communication skills, we’ll get better at understanding how our input affects the quality of GenAI’s output.
For starters, let’s simply play. If you have felt (understandably) resistant toward engaging with GenAI tools, or don’t even know where to begin, consider trying the following:
First, choose one GenAI platform to experiment with — for a few get-to-know-you exercises. Some options include:
- ChatGPT (the free version accessible through this link)
- Claude AI (accessible through this link)
- Google Gemini (accessible through this link)
- Microsoft Bing CoPilot (accessible through this link)
Next, create a free account. Then, try some of these “prompts”:
Prompt #1: Ask the chatbot to do a task.
- Example: Please write me a haiku about [insert an activity you enjoy].
- Example: Please give me five inspirational quotes about [insert an activity you enjoy].
- Example: Please suggest three gift ideas for [describe the recipient of your gift].
Prompt #2: Ask the chatbot to perform a task in the voice or style of a “persona.”
- Example: Please write [an imaginary but motivational paragraph] about [insert topic] in the voice or style of [insert name of a famous person you admire]. Change the “persona” and see what happens!
Prompt #3: Try a prompting technique called “tree-of-thought” (a debate on a topic from three or more points of view).
- Example: Please demonstrate a “tree-of-thought” debate among three people about [insert topic and specify whether you want one of the three people to hold a certain viewpoint].
Prompt #4: Ask the chatbot to outline a 3-day, 10-day, 20-day, or 30-day step-by-step plan for you to explore a new skill, or accomplish a certain goal, or change a particular habit.
(More Advanced) Prompt #5: Ask the chatbot to write about a topic you are interested in or working on — giving it these parameters:
- Give the chatbot a “role” to play (e.g., “You are a teacher/lawyer/parent/HR Director/leader of an organization, etc.”)
- Identify the task you want the chatbot to perform
- Describe the writing tone or style you want the chatbot to adopt
- Describe the format of the desired output (e.g., “Please write one sentence/one paragraph/a blog/bullet points/a tagline/a bumper sticker slogan, etc.”)
- Consider giving the chatbot an example of the output you’d like to see
This new technology certainly can feel disruptive and disorienting. But if we embrace it with a curious spirit, GenAI just might become a reliable creative companion, a motivational daily messenger we need, and a much-needed encourager along our journey toward leveling up our professional and personal pursuits.
**Professor Heidi K. Brown is Associate Dean for Upper Level Writing at New York Law School. She teaches legal writing and designs workshops, courses, and curricula around “writer identity formation,” including incorporating GenAI tools into writers’ workflow. For more, check out www.theflourishinglawyer.org or email Heidi at heidi.brown@nyls.edu.
