Will (Chat)GPT Improve the Way We Use the Internet?

Nowadays, it seems like we’re so quick to jump on trends and buzzwords, with one of the most recent being Chat GPT. We’ve had our fun and found much-needed answers, now it’s time to look at its long-term benefits and potential.
How can GPT improve the way we use the internet? And even beyond that, where can we apply this technology to add to our everyday life?

Reading Time: 4 minutes

gpt technology

Illustration: Milica Mijajlovic

What is GPT technology?   

You may have read somewhere that the ever-popular ChatGPT actually uses GPT-3. It’s okay if you had no idea what it meant. But the fear of missing out (FOMO) is so present in Web 3.0 that we all just nod our heads when we come across a new term like this. 

Here’s your chance to actually understand what GPT technology is, and what’s different with each generation. 

It may seem contradictory at first, but we’ll start with the latest model. 

GPT-3, as well as the previous versions of the original tool, is developed by OpenAI (the creator of ChatGPT). The name itself stands for Generative Pre-trained Transformer 3, which insinuates that it’s able to generate text by pre-fed data. Around 570GB of information, including Wikipedia content, has been fed into this tool so it can be able to carry out requested tasks. 

With that in mind, it’s useful to mention ChatGPT doesn’t have info about events and people after 2021, but its developers claim to work on expanding its knowledge. 

What’s more, OpenAI has an open call for any developer who wants to challenge this technology and explore its potential. 

From here, we can go on and say the previous generations of GPT-3 relied on the same technology but were pre-trained at a smaller scale. For example, GPT-2 was fed with “only” 40GB of internet text – which seemed fascinating at the time. 

However, the original model was never released due to concerns regarding malicious applications of the tool. Instead, OpenAI launched a “light” version of a smaller model. 

How can GPT improve our online experience?   

Experts have already tried to brainstorm ideas for how GPT could improve what we already do on the internet. 

For example, we’re already using chatbots, voice assistants and language translation but, by introducing GPT, these natural language processing (NLP) algorithms can become flawlessly accurate. Why not go so far as say that GPT was the missing piece for NLPs to reach the next level? 

Now, as a writer, I find it hard to believe that GPT technology has the potential to create valuable content and wipe content writers off the Earth. But I can see how it can happen with businesses that don’t focus as much on original content. That way, they would still have decent blog posts to fill in the space on the website and seem more professional. 

Later on, we’ll expand on how GPT can be used for creating written content not only for digital platforms but in education or administration as well. But before that, we need to emphasize that GPT-generated text has quite a few shortcomings that might not be noticed at first. Still, if it’s something important, you probably shouldn’t risk looking like a frivolous amateur. 

On a brighter note, one of the strongest impressions that ChatGPT left on users is its personalized user experience. You really feel like its responses are curated specially for you and that you have a special bond with this chatbot (now all the romantic sci-fi movies don’t seem as bizarre, do they?). 

What’s more, this experience can be brought way beyond the initial expectations of OpenAI’s test project. Imagine if you could be able to browse the internet and have the same feeling of it being your personalized space, giving you the exact recommendations you need. That’s definitely a possible scenario since this machine learning model could be used to analyze your browsing history and generate recommendations based on your interests.  

Last but not least, GPT-3 is a language prediction model. It doesn’t only refer to predicting the next useful information for the user, but it can also have a great impact on market research, financial analysis, and much more.  

However, as the CEO of OpenAI Sam Altman pointed out, GPT-3 is offering just a glimpse into how AI will change the world. 

Can GPT-3 be used in everyday life?   

Since GPT-3 can create anything that includes a language structure, its application can be widely considered in the offline world as well. We’ll give you a few examples. 

If you’ve recently spent some time on TikTok or Twitter, you’ve probably stumbled upon a viral video of using ChatGPT to write insurance denials. 

Namely, MC Clifford Stermer used ChatGPT to write a letter to a medical insurance company and explain why a patient with systemic sclerosis should be approved for an echocardiogram. He ended up very satisfied with the chatbot’s response. 

What’s important about this example is that it shows how GPT-3 could be used to save time in healthcare, by assisting with regular administrative tasks that can be automated. It can also be used to generate more patient-friendly descriptions of procedures or post-procedure instructions. 

However, considering that ChatGPT has many limitations, it’s advised to review and/or edit the text before sending it to someone. 

Another example shows that this chatbot can be used to write out summaries. As one Twitter user explained, he recorded the office meeting, used Google’s Speech-to-Text to convert it, and fed this transcript to ChatGPT to summarize it. The result was, as you can read in the comments, jaw-dropping. 

And the list goes on. 

Besides mentioned examples, there’s a whole grey area in using ChatGPT in education. Apparently, teachers all around the world are having headaches because students are excessively using this chatbot to write essays. The worst part is, GPT-3-generated text oftentimes doesn’t show up as plagiarised (yet) so it’s almost impossible to prove that a particular student has cheated, even though his previous grades are nothing compared to the ones he’s getting, thanks to ChatGPT. 

But we’ll leave that for another occasion because we’re definitely not done with the potential of GPT-3 technology. 

A journalist by day and a podcaster by night. She's not writing to impress but to be understood.