Chat Generative Pre-Trained Transformer, colloquially known as ChatGPT is an internet-based computer program designed to take written human language input and create an essay, article or summary of that input, based on the user's needs. ChatGPT was created by OpenAI, a software company based in San Francisco. Anyone with a valid email address can use ChatGPT, provided they are over the age of 18. There are two versions of the software available, a free and a paid version.
DISCLOSURES: The author is not a computer software engineer and has no affiliation with OpenAI or ChatGPT. You may utilize the ChatGPT program after making an account with OpenAI on the OpenAI website.
Although chat-based natural language programs have only been gaining popularity for the last 20 years or so with the advent of the Loebner Prize in 2006, one could argue the modern concept of computers imitating human language took hold in human consciousness with the proposal of the "imitation game" by Alan Turing in 1950. Since then, the type of challenges computers and computer programs competed in became known as Turing tests. A computer/computer program passes a Turing test when the actions taken by the computer/program are indistinguishable from the actions taken by a human, judged by a panel of people. Most commonly today, the action judged is of written, natural language communication, but the Turing test could be applied to any behavior or action taken by humans, including: spoken language, driving a vehicle, or emotional expression, to name a few.
But how does OpenAI's ChatGPT benefit humanity? It could be troublesome having this program freely available to the public given the human tendency to want to put in as minimal effort as possible when acting as an end user of software programming.
One area of concern would be that human natural language learners, i.e. children, may use this program/platform for plagiarism purposes. Though OpenAI does have a software tool to detect use of ChatGPT.
Some questions that need answering include:
Does using ChatGPT make human communication easier or more consistent?
If the answer is "Yes," will those who choose not to use it or don't have meaningful access to ChatGPT be at a disadvantage?
It is possible that much of the input ChatGPT receives will be repetitive information, how does it work to improve the use of "natural language" to improve human comprehension?
For the last question, at least, we have a partial answer. The best information available currently (as of the writing of this article OpenAI has not disclosed how they incorporate new data publicly) is that the creators of ChatGPT likely perform Supervised Learning and Reinforcement Learning from Human Feedback.
So, much like a human learning a language for the first time, the creators and programers responsible for ChatGPT are continually modifying the coding to better adapt ChatGPT to human style communication.
It seems though, when we really think about it, it's the messiness of communication that might be one of the most entertaining aspects of being human, as any parent can attest to when teaching their child to speak. Kids say the darnedest things.
And although, fundamentally, language is a work of human "cognitive engineering" it might very well be that the best writings of our generation -- and successive generations -- will be set written by the muse-in-the-machine computer programs and their small army of human software teachers.
An excerpt from InstructGPT, a precursor to Chat GPT:
Don't programs like ChatGPT say the darnedest things?
There is still a long road ahead to bring up a socially-productive Chatbot. And sure, programs like ChatGPT will say some interesting things from time to time. But the road is wide for computers to captivate audiences with the capability of software programs to create a natural language Chatbot which aims to enrich our lives and improve the human experience.