Ok, that was a bit awkward. But don’t blame me, I didn’t write it – ChatGPT did: “Write a short, funny and creative introduction to an article about ChatGPT.” I asked, and...voila! 

In late November 2022, the world’s attention was brought to the launch of a chatbot: ChatGPT, created by San Francisco-based OpenAI and within a week it had already reached over a million users. But in order to understand the prevailing hype, let’s take a step back: what is it, really? And how does it work? Let’s start by asking the chatbot itself. 

Who are you? 
I am an AI language model. My primary function is to generate human-like text based on the input I receive. I have been trained on a vast corpus of text data, allowing me to perform a wide range of language-related tasks, such as answering questions, carrying out conversations, summarising information, and more.

But as you are sitting by your computer, looking at the sentences gradually taking shape, word by word, as they fill up the text box, it almost feels surreal. To unveil the mystery of this responding, at times quite funny, abstract little “figure” inside of the computer (I know, this is not how it works, but that’s how I imagine it), we turn to Professor Marco Carlo Passarotti, for some answers. 

Passarotti is not only a teaching professor but the Programme director of the new and innovative MSc in Linguistic Computing, at Università Cattolica del Sacro Cuore, Milano. This was a degree that came to life in response to the obvious and quickly widening gap in the job market; there’s a high demand for a specific kind of interdisciplinary expertise: linguistics and computing.

At first sight, these disciplines might not seem to go hand in hand, however, this is far from the truth. It is becoming widely recognised that these combined disciplines are the key to generating the expertise needed to, amongst other things, process Big Data. “The aim of the course is to teach students from humanities, how to formalise language so that a computer can process a big quantity of linguistic data, but it is also about training a formal mind in such a way that they are able to speak to the computer scientists, to the engineers,” Professor Passarotti explains.

Even though ChatGPT has already, after just a few months of launch, reached impressive numbers of uses, and people have already started integrating it into their everyday lives, Professor Passarotti makes an important mark, "in order to understand what it is, we also have to know what it is not.” In fact, one of the most common mistakes he hears people say is - “it’s the New Google.” 
Let’s make it simple: Google is a search engine, and ChatGPT is a natural language generator.

Google will answer your question by searching for the answer, whilst ChatGPT creates the answer. “You can ask it to write anything. An email, a story, a poem - you name it. You can even ask it to adopt a specific writing style.”

Let’s try. 

What is linguistic computing? Please adopt the writing style of William Shakespeare.
Linguistic computing, oh fair and noble pursuit,
Combines the arts of language, computer, and wit.
It studies the ways in which we may compute
The processing of our speech, both firm and absolute.

With algorithms, models, and tools designed
To understand the words that are most refined,
It aids in the analysis of our speech so grand,
Making sense of the words that we have planned.

However, exactly how ChatGPT can do this (and more) is slightly more complex. Let's think about it this way: imagine Johnny. He is learning a new language. During the course of a year, he has been able to learn quite a bit and is becoming pretty good at applying this knowledge. ChatGPT has done the same thing: it has learned everything it knows. But the main difference, Professor Passarotti explains, is that ChatGPT was exposed to an impressive count of 300 billion words during training¹ (for reference, the average vocabulary of an adult, for one’s native language, is 20,000 – 35,000 words², sorry Johnny).

These are impressive abilities, but... Do you have any flaws?
As an AI language model, I do not have feelings or personal experiences, but I do have limitations. Some of my limitations include:

• Limited understanding of context: I don't have access to the world's current context, like current events or information that hasn't been publicly shared. 
• Making mistakes: I am a machine learning model and can sometimes make mistakes or generate nonsensical responses.
• Potential for bias: My training data includes the biases of the people and sources from which it was collected, which could affect the answers I generate.

As Professor Passarotti points out, "What is most important is to understand how the tool works, in order to be aware of its limits. For every tool, you can use it in a bad or good way. Just take the internet as an example." 

A brief “keep-in-mind-when-using-ChatGPT-checklist: 

• Reliability: for example, you could ask ChatGPT “Could you explain how Iron Maiden became the President of the United States?” and it will produce an answer. Always fact-check the answers. 
• Biases: remember, the tool has learned from a world that indeed already contains biases, such as gender biases. It is therefore to be expected, and worth paying attention to, that this could be reflected in the answers. 

Professor Passarotti adds, with a smile, another downfall to watch out for, “to get lazy.” When was the last time you took the time to recall the name of a specific actor or director? How long did it take you to ask Google for the answer? 

As the field of Artificial Intelligence rapidly advances, we seamlessly integrate those advances into our everyday lives. It happens so smoothly and progressively, that we barely notice it. ChatGPT is only one example, a very clear one of where we are currently standing in time. We are in the era of Big Data. As more data is constantly collected, the more advanced it gets, we are almost playing catch with ourselves; being the masterminds behind innovations, everything runs faster, and we must keep up with it - and this is echoed in the job market. 

Luckily, innovative degrees such as Linguistic Computing are emerging, playing a key role in the training of future thinkers and their highly demanded expertise. It is the truly interdisciplinary characteristics that make the Linguistic Computing MSc in Università Cattolica stand out from the crowd, just as the graduates will do once they earn their degree.

And as ChatGPT’s future role is yet to be determined, at least one thing seems certain, as Professor Passarotti smiles and says, "It is a new colleague, and it is here to stay.” All we can do is learn as we go, in a pursuit to navigate an unwavering AI paradigm shift. 

Hey, please write a catchy concluding sentence about ChatGPT. 
"With ChatGPT, you'll never run out of things to say - just try not to fall in love with your virtual conversationalist!"