Tech Tuesdays #1 - ChatGPT and AI
That's ^ the image that comes up when you search for 'AI' in Wix's free image collection btw. Fitting. A lot of people are terrified of AI. The NYC Education Department certainly is.
Before we jump in, just wanted to say 'hello.' Welcome to the first ever Tech Tuesday! I'm going to use Tuesdays to rant about random tech things, and today is no exception.
Today, I'm going to talk about my views on ChatGPT and AI, as inspired by Fred Wilson's post here. For clarity - I don't know Fred personally (but hope to, one day! I hope he invests in one of my ideas at some point tbh). I do however follow his blog and get his emails. Lots of great insight in those posts so I recommend checking him out.
My introduction to ChatGPT and OpenAI was actually in a work Slack a few weeks ago (I already knew about it but hadn't played around with it yet). Our CTO asked it to write a humorous bio for a few new employees. He also asked for one about our CFO, and it was pretty good. I asked OpenAI for one of myself, and unfortunately it was subpar. It couldn't find me, and it ended up writing a bunch of weak bios for other Indranil Mitras - even when I tried to get it to reference my LinkedIn profile directly.
AI is only really as good as its training data, so maybe simply because I fed it info about me, if I were to return to the bot in a few weeks, it would be able to write about me. Maybe I also need to be more famous to become the #1 Indranil Mitra (maybe this blog will help a bit).
Fred Wilson discusses in his post, referencing the Chalkbeat article, that the NYC school system banned the use of ChatGPT, especially for essay writing. I imagine that with easily accessible online data, ChatGPT would write a pretty solid essay (realizing now that I should have used it for this post). I've found that while the bot isn't always perfectly on target in what it is referencing, the writing itself is remarkably human. It really seems like a person wrote it.
One may argue that it's inevitable that AI will take over writing, and I have actually seen syndicated articles popping up in large publications that note they were written by AI. Why pay a reporter to summarize an existing article when their time is better spent creating brand new content? Similarly, it seems inevitable that students will leverage tools like ChatGPT for their papers. Who wasn't looking for a shortcut as a student (hello, WolframAlpha)? Outside of writing, it also seems inevitable that AI will start to take on tasks like scheduling and making reservations. Google already has a feature where it can call to book an appointment for you. It's not super human sounding, but based on the progress of ChatGPT, I feel we're not far away from a realistic virtual personal assistant.
AI has a ton of practical applications and maybe it can get to a point where it takes on menial tasks that we don't want to do, and enables us to instead focus time on things we do want to do (sports, arts, etc).
So how do I feel about AI? Despite the advances, I'm ambivalent. Not necessarily in a terrified, Skynet, way. But more so in a 'I see the benefits but feel we're very far away' type of way. Firstly, ChatGPT can't even write a bio about me, even when prompted with source material. I don't feel like it's going to take my finance job tomorrow. Second, there are social limitations to what AI will be allowed to do. While AI is creating beautiful artwork now, I don't actually see it taking over all arts. Things like filmmaking and photography will remain human-led. After all, we're saving time and money in order to do things like filmmaking and photography. If AI takes those, and we get stuck just doing our jobs, people will riot and throw the whole thing out.
In regards to students using ChatGPT to write papers, while I agree that it may be good to have them learn to use it as a practical skill, they should also actually write the papers. Communication skills are absolutely key to personal development, and if we let AI do everything, we'll end up in a world where tv/movies have to be written by robots - simply because we'll forget how to write things ourselves.
So, while I think AI has a lot of practical uses and I'm excited to see how far it has come, I think there are limits to it, and even limits we should put on it. We should focus AI on things like data processing and medical diagnosis support. We can use it for things like scheduling appointments for us. We should not however encourage it to replace artistic endeavors (or upload us to the matrix).
Kids, write your essays. Hopefully they're more coherent than my blog posts.