33
Michael Hallissy and John Hurley
History
In late November 2022, the world was introduced to ChatGPT, an Artificial Intelligence (AI)-based chatbot system based on a language model called GPT-3.5. It uses natural language processing (NLP) to generate conversations1. ChatGPT was the latest in a series of such tools but, unlike previous tools, it has captured public interest and imagination. It had more than a million users within a week of its launch because of its ability to generate human-like text and its perceived implications and potential use in education, the workplace and everyday life. ChatGPT can answer questions and assist you with tasks such as composing emails, essays and code2.
GPT-3 (Generative Pre-trained Transformer-3) is a Large Language Model (LLM) that was trained through deep learning on a vast quantity of data (499 billion data points – 800GB of data), a model some one hundred times larger than any previous models3. ChatGPT enables human-like conversations with this model. A large language model is trained to predict the next word in a sentence, somewhat in the same way that a text-message auto-complete-feature works. It continues this process to generate more text.
Since its arrival there have been many developments in relation to the creation and development of generative AI tools. We have captured some of these in the graphic above. Prior to November 2022, few people knew what generative AI was. All of a sudden, terms such as ChatGPT became everyday terms for many – particularly those in education. The indications are that these technologies will continue to develop and become more useful to all in education.
Generative AI – What are they?
Generative AI systems create new content in the form of images, text, audio, video, and more:
-
- Images: generative AI can create new images based on existing ones, such as creating a new portrait based on a person’s face, or a new landscape based on existing scenery;
- Text: generative AI can be used to write news articles, poetry and even scripts. It can translate text from one language to another;
- Audio: generative AI can generate new music tracks, sound effects and voice acting4.
The list of possibilities continues to grow, but let’s start by asking some generative AI tools how ChatGPT got its name.
But what is in a name?
Generative AI tools have multiple uses, so let’s give a number of tools – Bing Chat, ChatGPT 3.5 and Google Bard the following prompt:
You are an expert in AI in Education. Imagine you are explaining what ChatGPT is and how it got its name to young children, aged between 10 to 14. Explain why it includes the term chat and show some examples of how it can be used in education. Write your output as a persuasive piece of writing that explains ChatGPT
Here is what Bing Chat created:
The name “ChatGPT” comes from the words “chat” and “GPT.” “Chat” means to talk with someone in a friendly and informal way. “GPT” stands for “Generative Pre-trained Transformer,” which is a fancy way of saying that ChatGPT is really good at creating new things like text, images, and sounds.
Here is what ChatGPT created:
Now, you might wonder why it’s called ChatGPT. Well, the ‘Chat’ part is because it loves to have conversations with people. You can type in a question or a topic, and it will do its best to give you an answer that makes sense. It’s kind of like talking to a friend who knows everything about, well, almost everything! Plus, it’s really good at chatting in different languages too, which makes it super helpful for kids all over the world.
The ‘GPT’ part is a bit more complicated. It stands for “Generative Pre-trained Transformer,” but you don’t need to remember that tongue-twister. Basically, it means ChatGPT learned a lot by reading tons of books, articles, and websites, just like you do when you’re studying. It’s like it went to school for a super long time and became an expert in all sorts of subjects!
Here is what Bard created:
ChatGPT is a type of artificial intelligence (AI) called a large language model. This means that it’s been trained on a massive dataset of text and code, which allows it to generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way.
So, why does ChatGPT include the term “chat”? Well, because it’s designed to have conversations with people! You can ask ChatGPT questions about anything, and it will do its best to answer them in a comprehensive and helpful way. You can even have fun with ChatGPT by asking it to write stories, poems, or even code.
These tools allow us to write prompts and then have a conversation (conversational AI) with a ChatBot. There are a growing list of ways, which we touch on in the textbook, how these tools can help teachers, students and administrators in education and training. And they allow you to have a conversation, so think of the ways that might help you or your learners – you can have a chat about a topic you need to research or to prepare for an interview or to generate ideas for a lesson or a paper. It is interesting to compare and contrast the outputs they each create and then pick the best pieces for your purpose.
These tools can ‘enhance’ or assist you to be more creative, productive and effective when you combine your human abilities with the abilities of the machine. So always remember to critically evaluate what it produces. So have some fun learning how to have conversations with these tools and see what they produce and if they are useful to you and your students.
Generative AI in Numbers5
The following graph6 shows the rise of ChatGPT between November 2022 and May 2023. It gained one million users in its first week after it was launched. It should be noted that Netflix took 3.5 years to reach a similar milestone, with Twitter taking two years7.
While ChatGPT reached its peak in May 2023, it continues to retain its popularity with other generative AI tools such as Bing Chat and Bard also gaining traction. In the case of Bing Chat it started to become more popular when Microsoft acquired ChatGPT in March 2023, while Bard continues to grow in popularity as well.
3 https://hai.stanford.edu/news/how-large-language-models-will-transform-science-society-and-ai
4 https://www.weforum.org/agenda/2023/02/generative-ai-explain-algorithms-work/
5 https://research.aimultiple.com/generative-ai-applications/
6 https://www.tooltester.com/en/blog/chatgpt-statistics/
7 https://bootcamp.uxdesign.cc/chatgpt-vs-bing-chat-which-is-better-2e46fa821d7d
