As I am with most things that generate a buzz overnight, I was a bit weary of GhatGPT. Of course, I recognized the impressiveness of its text-generation capabilities, as it was a huge jump from what I’d seen before.

But there was this nagging thought at the back of my mind that just wouldn’t let me get too excited about GhatGPT.

What was it?

Was it fear of the unknown?

Most people tend to fear novelty; it’s how we are wired. We find a lack of information unsettling because we don’t know what to expect or how to react. But I haven’t been afraid of progress before, so why would this start nagging me now?

Was I not understanding the technology?

I might not be familiar with the specifics of the implementation, but I like to think I know enough to understand what’s happening behind the scenes. Many have called ChatGPT merely a probabilistic model or “autocomplete on steroids”, and I have to agree. So I figured it wasn’t the technology that was bothering me.

So maybe I was fearing how this new technology would be used.

The hype around it ensured that even the most removed from the world of tech or AI would hear of it. And while most were praising the human-like responses of the chatbot, a lot fewer were warning about the pitfalls.

There is so much information out there, it’s sometimes difficult to navigate. So having something help with a quick(er) answer or a summary is great, as long as we keep in mind that we always need to verify the validity of the information. However human-like the responses of ChatGPT are, there is no guarantee of correctness. The chatbot just doesn’t know what’s true. And more dangerously, it doesn’t let you know it. It will provide an answer no matter what.

Add to this the fact that we don’t really know what data was used to train ChatGPT and I find even more incentive to be on the untrusting side. You wouldn’t ask a painter to help with your code, and you wouldn’t ask a software engineer to put in your windows. So why trust something trained even on random people’s text on the internet?

All in all, like any technology, ChatGPT can be used for good or for bad. I find it has immense potential to aide in creative tasks, but I strongly believe we shouldn’t use it as anything other than a tool. And we should always, always check the facts by ourselves.