Higher Ed's Obsession of the Week: ChatGPT

Let the hang-wringing begin in earnest. (Note: this post 100% human-sourced, non-GMO, and certified organic. Also, mostly gluten-free.)

If you’re a regular reader of The Highlight Zone, but you haven’t yet subscribed, consider plugging your email address now so I know you’re reading. I would really appreciate it! Or, even better, go ahead and become a paid subscriber or founding member (‘tis the season for giving, after all), and you get the satisfaction of knowing that you are supporting my work.

At least this gives #AcademicTwitter a break from obsessing over whatever fresh hell Elon Musk is waiting to unleash on those poor Twitter employees so they can instead focus their energies on obsessing over another of his investment projects. (Musk stepped down from the board of OpenAI in 2018 citing conflict of interest concerns with Tesla, but remains an investor.)

On November 30, ChatGPT (Generative Pre-trained Transformer) debuted. This is a new version of existing software, and from what I’ve been able to learn, it avoids many of the pitfalls and bugs of its predecessor, GPT-3.5. There are shortcomings: it has a tendency to produce inaccurate responses, and it’s sometimes picky about how users phrase their questions or prompts, but generally speaking it’s a game-changer. Harvard Business Review described ChatGPT as a “tipping point” for AI. Users can converse with it in a way that feels natural, ask for essays written in the style of Shakespeare or Fitzgerald or Trump or even themselves (see below), and even have it write serious stuff like contracts and other legal documents. I’ve only played around with it for a few days, but I have noticed that it has a penchant for writing five-paragraph essays: “There are benefits and drawbacks to X. Here is a benefit, and here is a another benefit. Here is a drawback, and yet another drawback. In conclusion…” That sort of thing.

But don’t get me wrong, it’s super cool. Seriously, check this out: