Share this...

By Trisha Prabhu

“Hi Trish, I’m wondering what personalized GPTs are. I know what ChatGPT is, but what are personalized GPTs?”

Hi there, and welcome back to another week of Ask Trish! I hope you’re all well and having a great start to the week. (For my friends in the US, wishing you and your families a very Happy Thanksgiving this Thursday!)

Thanks so much to this week’s question-er for the fantastic question. AI truly is all the rage/constantly in the news these days and has been the subject of several of my most recent posts (check out my last post on the UK AI Safety Summit here!). It seems like there’s a new AI-related development everyday…which can make it hard to stay up-to-date. But that’s what I’m here for! In this week’s post, I’ll break down one of the most recent AI-related headlines we’ve seen: the release of personalized GPTs. I’ll briefly take y’all through 1) what personalized GPTs are, 2) why commentators and observers believe this development matters, and 3) what, if anything, we ought to be concerned about when it comes to personalized GPTs. I’ll leave an overall assessment of this new technology to you all. 

Sound like a plan? Let’s get into it:

First of all…what the heck are “personalized GPTs”? What does that even mean? Personalized GPTs are exactly what they sound like: AI chatbots that you, yourself can create and personalize. (Think of ChatGPT – but tailored to you!) In particular, you can personalize these chatbots to complete specific tasks for you; you might, for instance, personalize an AI chatbot to serve as a math coach for you! And that’s just one use case; of course, the possibilities are endless. If you’re thinking…Yeah, but Trish, I don’t really know how to code, so I don’t think I could personalize a chatbot, don’t worry – OpenAI (the organization behind ChatGPT) has built this new service such that you can personalize with no coding required. Perhaps most interestingly, not only can you now create a custom GPT, you can actually “plug it in” to other parts of your digital world, e.g. your calendar, Slack, etc., and enable the chatbot to perform tasks for you on these services. (Amazing…and a little scary, right?) OpenAI’s larger vision for these personalized chatbots, which it calls “GPTs,” is to create a “GPT store” of various AI agents that you, as a user, can purchase or you, as a creator, can help build. These agents may ultimately do everything from calendar scheduling to homework help to even more complex tasks, like helping you purchase Christmas gifts for your family members.

Okay, great. So we now know what these GPTs are…but why do they matter? Well, as New York Times contributor Kevin Roose recently outlined in this article, you might think of the era before personalized GPTs as an era in which AI could talk about things (cool, but the applications are limited)…but an era post-personalized chatbots is an era in which AI can actually do things for you. That’s a world in which AI has the potential to dramatically reshape our lives, possibly in ways that we can’t even imagine now. Today, it’s likely that you only consciously interact with AI in limited ways, at limited times; in a world in which AI is regularly performing tasks for you, AI will be actively and constantly creating the circumstances in which you live your lives. And it’s not just us – imagine, as Roose suggests we do, what this type of AI could do for companies, governments, and large organizations. The effects are likely to be game-changing. Personalized GPTs, then, really reflect a new phase for AI. Of course, we’re not quite there yet…as Roose describes, the personalized GPTs he’s worked with are imperfect. Even so, their potential is immense – and that’s huge.

You may now be wondering…okay, but what’s the downside? Could anything go wrong? Of course – as is the case with any emerging technology – we always have reason to be cautious, to carefully probe the potential issues the technology in question may present. In this case, right off the bat, it’s obvious that one potential issue is that faulty AI may make inaccurate or bad decisions. AI saying the wrong thing is one thing; AI doing the wrong thing is an entirely different thing. If, because of biases or poor development, AI begins to make decisions in ways that systematically harm certain groups, it may inadvertently reinforce inequalities or oppress certain groups. Another potential issue is the way that this type of AI is going to disrupt existing jobs/labor. If AI starts doing a ton of tasks for us…what are we going to do? It seems like a silly question, but in fact, it’s a very relevant one, one that lots of folks are thinking about now. Where do humans fit in this new world? Finally, a lot of the issues we see in Web 2.0 – like fraud and harassment, for instance – may manifest in this new world, possibly in ways that are more intrusive and extreme. Tackling these issues, then, is likely to be crucially important.

As always, thank you so much for reading along! I hope that for many of you, this was an informative, valuable first look at personalized GPTs. (Now that you have a sense of what they are, I encourage you to do additional research about them and learn more!) If you’re looking for answers to more AI questions – or perhaps seemingly more simple (but certainly still important!) questions – please join the conversation! Share any Internet-related questions, thoughts, or perspectives here (and your topic just might be featured in an upcoming TikTok/blog post!). Don’t forget: anything you’re experiencing or wondering about is 100% valid (and there’s a very good chance that other folks in our community have the same topics on their mind!). 💙Oh, and one last thing – don’t forget that if you see an Ask Trish video you like, you should give it a like and share it on your social media. Let’s spread the #AskTrish word this week (and every week)!

Have a great rest of your week (and again, Happy Thanksgiving!),


Share this...