04/15/2024
Today's post is an opinion on LLMs from our consultant Milton!
"My thoughts on AI/LLM platforms are my own opinions, and don't necessarily reflect the views of BC nor the BCWC.
After playing around with ChatGPT and participating in AI trainings, I honestly do not see much benefit in using AI/LLMs in Writing Centers. Currently, there are is a lot of discourse around their "utility." However, the truth is that LLMs are still limited in their capabilities.
AI platforms/LLMs (primarily ChatGPT) tend to “hallucinate,” which is a new term used to describe when LLMs generate deceptive information that seems legitimate. In my experience, ChatGPT has repeatedly cited fake sources, even when correcting my instructions. This is dangerous as it may perpetuate the societal issue of misinformation, and can even lead to serious academic consequences if a student were to use ChatGPT as a research tool. So, I’ll stick with doing scholarly research the "old school" way.
I’m also aware of the ethical challenges presented by AI/LLMs and their popularity. More specifically, these platforms extract data and work from writers, artists, musicians, etc. with little or no consent. This issue is hardly discussed unfortunately because it's overshadowed by the admirations of AI for its “convenience” and “novelty.” Thus, I believe those conversations need to pause, and this issue should be taken more seriously.
The only benefit I will acknowledge is that ChatGPT can be used for brainstorming. For example, if a writer experiences writer’s block, they can consult ChatGPT for additional points to write. However, I’m a senior consultant with several years of experience in writing and tutoring, and a degree in Communication. ChatGPT, or any other LLM, is not a gamechanging resource for me. My own expertise is far more reliable and authentic; nothing that any programmer can replicate. Although LLMs might be faster, they are certainly not better than me."
Image description: A meme image that portrays Kermit the frog standing face-to-face with his dark ego. Normal Kermit's caption is "find 10 scholarly articles" and Dark Kermit's caption, representing ChatGPT, is "make them up."