Reevaluating My Place on the Curve

Looking back on my initial thoughts at the start of LDT 511, I pendulated between feeling like an early majority versus a late majority adopter. Now that the course is coming to a close, I feel more confident identifying as an early majority adopter when it comes to new tech, as a result of the experiences I’ve had over the last several weeks.

Somehow, though, I still feel obligated to identify as a late majority adopter. In my current role at work, I will not be using emergent technology any time in the near future. I’ve enjoyed thinking through the way VR and AI could be introduced in our college, but at this point, it’s back to established tech for me until the higher-ups decide it’s time (to pay) for a change.

Regardless, this course provided me with some pivotal moments that helped change my perspective of  emergent technology. One primary example is how this course facilitated my first in-depth interaction with ChatGPT, which has been key in helping me gain entrĂ©e into the nebulous world of generative AI.

A fear I had prior to this course is that I would unknowingly accept false information from ChatGPT, regardless of the ever-present warning below the chat box: ChatGPT can make mistakes. Consider checking important information. But it’s not the important information that I’m worried I’ll miss – it’s the minor details that I worry will in passing make me look misinformed and ignorant, diluting what may otherwise be a solid and convincing monologue. I don’t know what I don’t know.

The design case in which we had to ‘partner’ with AI to craft a children’s folktale was the first time I had experimented in depth with ChatGPT, and it helped calm my fears that this chatbot is somehow omniscient and thus – I don’t know how else to say this – smarter than I am. Based on the requirements of the assignment, I asked ChatGPT it to “Generate a 400 word children's folk tale about a polar bear living in the Arctic region who is dealing with the effects of climate change.” In response, ChatGPT generated a solidly boring story about a Polar bear it named Frosty.

I was going to leave it be, but halfway through the story, ChatGPT wrote that Frosty the Polar Bear planted trees to help stop climate change. I stopped reading. Planting trees? To stop climate change… in the Arctic? Oh, absolutely not. Sure enough, I did some research about this minor detail and discovered my intuition was correct – tree growth in the Arctic is actually a sign of climate change, not a way to mitigate it (Diaz, 2022).

So, I told ChatGPT, “Try again but don’t have them plant trees.” Not my best prompt, sure, but instead of removing the reference to trees, ChatGPT added a statement that “the key [to fighting climate change] lies not in planting trees.” What?!. 

“DON’T MENTION TREES,” I yelled at the chatbot. 

“And so, the Arctic flourished, not by planting trees,” ChatGPT spat back at me in the conclusion to its regenerated story. 

I doubled down. “DO NOT INCLUDE THE WORD ‘TREE,’” I screamed. 

Its response? Another regenerated story that still included that exact same line in the conclusion. I gave up and reset the thread, giving it a prompt about ‘melting ice’ rather than ‘climate change’. Trees were not mentioned again.

Following this experience, I tried to get ChatGPT to help me craft a recombinant technology name for the next unit’s white paper. It was an odd experience – the inverse of what I was expecting. ChatGPT generated a handful of bland names for my technology, no matter how much detail I provided. It could not stop suggesting names like “MarkBiz” and “SimPlus” and “BizPro”. I learned later that another colleague encountered the exact same problem. In the moment, I was so frustrated with ChatGPT that I actually think it incentivized me to come up with a name on my own, which is what I was hoping to do all along. I found it ironic that in this interaction, instead of guiding me in the right direction, ChatGPT simply guided me away from the wrong direction.

And yet, from these pivotal experiences with key technology in the course, I found my attitude towards new tech has evolved differently than I anticipated it would – and unexpectedly, this has made me more willing to work with it.

For instance, though I did not have the opportunity to use VR directly in this course, I engaged in some thought experiments about it. I learned about tech designed specifically to enhance learning experiences for business students, and I chewed my way through (possibly Orwellian) ideas about using VR to inhibit academic dishonesty. My approach to emergent tech in the context of academic integrity has felt like an offshoot of the adage that “AI won’t take your job, people will” (Palmer, 2019). In my case: even if we don’t use emergent tech to prevent cheating, students will use it to cheat. All of a sudden, VR began to feel like a viable tool rather than a sci-fi gimmick. 

Circling back to generative AI, I get it now. I’m no longer worried that it’s a daunting black box of omniscient wisdom. I know how to leverage it to do what I want, and I’m learning its voice – which, so far, is just not great. It writes like a well-spoken student who skims readings and chunks together lukewarm essays. (My mother, a high school English teacher, jumped on ChatGPT when it was released to see what it would do with her essay prompts. Her conclusion? Just as she can recognize the voices of her students in their papers, she can recognize ChatGPT’s "voice" – and it’s not passing her Turing test.)

That said, while in this moment I am amused at how clunky the currently available generative AI technology is, Udell and Woodill (2019) are right to warn us that the 20th century “wait and see” approach to emerging technologies is not viable with today’s tsunami of new tech. I have no doubt that someday, AI-generated content will become indistinguishable from human-generated content. Despite my reluctance to accept change, this course has demonstrated to me that it’s worth taking the time to lean into emergent tech, rather than back away from it – or as Udell and Woodill put it, to “enjoy the shock of the new” (2019, p. 137).

References

Diaz, C. (2022, August 17). Spruce trees have arrived in the Arctic a century earlier than expected. Here's why. World Economic Forum. https://www.weforum.org/agenda/2022/08/climate-change-spruce-trees-arctic-century-earlier/

Palmer, S. (2019, November 16). AI won’t take your job, people will. Shelly Palmer. https://shellypalmer.com/2019/11/ai-wont-take-job-people-will/

Udell, C., & Woodill, G. (2019). Shock of the new: The challenge and promise of emerging technologies. American Society for Training and Development.

Comments