What is ChatGPT and how does it work?

By now you must have heard of ChatGPT, the text-based AI ‘search engine’ people are using to look up everything, from recipes to relationship advice.

If you have not, you have been living under a very large rock. ChatGPT (Generative Pre-Trained Transformer) is a chat bot by US start-up OpenAI.

It was launched on November 30 and is capable of answering a variety of questions. OpenAI was cofounded in 2015 in San Francisco by billionaire tech mogul Elon Musk, who left the business three years later, in 2018.

AFP reported that the start-up received $1 billion from Microsoft in 2019, growing to become famous for its automated creation software: GPT-3 for text generation and DALL- E for image generation.

With ChatGPT, OpenAI says it uses the GPT3 AI platform which houses some of the world’s largest language models, using both supervised and unsupervised AI learning techniques.

How does it work?

To use ChatGPT, one creates an account using their email address from the chat.openai.com web platform.

Similar to regular search engines, you type a question and the chatbot generates answers which are organized in a thread and saved for reference.

ChatGPT remembers all previous conversations and queries you entered and makes a reference to them in subsequent answers to drive a point home. For instance, it will give an answer like, “As I have said,…”

Also worth noting is that the chatbot filters its responses for inappropriate prompts, which are immediately identified and dismissed.

ChatGPT even develops codes that can be embedded in websites and can explain scientific concepts, write scenes for a play, academic papers and write poems.

You can choose to have it generate forms for your answers, although it is Text-based and thus cannot generate questions in other document formats such as .docx or .exe.

Although the service is currently free to the public, Forbes reports that it will be monetized later.

Amid the rave reviews (The New York Times called it the world’s- best AI Chat bot ever released, while The Guardian said it generates human-like text), its efficiency has however raised concerns among users and experts.

People are worried about its influence on things like human labour, cyber-security and academia.

Even Musk himself described it as “scary good”, tweeting on December 3, 2022; “ChatGPT is scary good. We are not far from dangerously strong AI.”