ChatGPT is not a person, despite how intelligent it may appear to be.

The recent release of OpenAI’s ChatGPT has generated buzz in the tech and online communities. The chatbot powered by artificial intelligence has shown the capacity to comprehend queries and prompts in everyday English and react in what appears to be similarly natural language. A computer programme called ChatGPT has been “taught” to respond to questions and commands in natural English. Almost 10 billion words and over 8 million documents made up the text set used to train ChatGPT. To further hone the questions and responses, ChatGPT was trained by human “coaches”.

Some people have found the resulting chatbot to be eye-opening and fascinating, while others have found it to be frightening and menacing. Any sufficiently sophisticated technology is indistinguishable from magic, as Arthur C. Clarke wrote in his widely cited 1962 essay, and this very advanced new technology has produced a response comparable to how people react to magic. It’s good to give some instances of what ChatGPT can achieve and what it currently CANNOT do to allay certain anxieties and appropriately address others.

Chat GPT’s capabilities
Let’s look at 3 things that this incredible new technology is capable of.

1. ChatGPT Can Communicate Clearly
   The ability of ChatGPT to comprehend human language, swiftly retrieve a large data set, and react in plain language can undoubtedly, in some ways, enhance the online search experience. Almost 30% of all internet traffic, according to a November 2022 Statista.com report, is directed to search engines. It is conceivable that ChatGPT could alter a sizable portion of internet usage. Asking a question on the internet and receiving a single straightforward response in our native tongue would be a significant improvement over the list of links stuffed with advertisements we currently encounter. ChatGPT can replace the need to visit numerous links and browse a variety of websites.
2. ChatGPT CAN Operate Rapidly
   Even vast amounts of content can be produced fast via ChatGPT. With the strength of its AI and the excellence of its language models, ChatGPT can without a doubt produce vast amounts of text quickly. No human can study information and provide pertinent content as quickly as ChatGPT. Let’s imagine the user requests 1,000 words on a specific subject, but ChatGPT’s initial response is only 250 words. When requested, ChatGPT will quickly generate more and more when given instructions, generating hundreds of words in a matter of seconds.
3. ChatGPT CAN Provide Details
   The explanations on ChatGPT are excellent. ChatGPT may assist in the simplification and clarification of these complicated issues, which can range from code to physics to ChatGPT itself (!). ChatGPT is an excellent tool for swiftly grasping concepts on a basic and superficial level because of its extensive database and the language skills it has been taught. You don’t know what the second law of thermodynamics is, do you? In a matter of seconds, ChatGPT can spit out a 36-word response that may be used as a quick reminder or as a reference without the need to click on any links.

Chat GPT’s Incapability
Despite how impressive this technology is, let’s be honest about what it cannot achieve.

1. The Truth Cannot Always Be Spoken by ChatGPT
   The ability to discover the truth may be the biggest issue plaguing internet searches (and possibly the internet as large). ChatGPT cannot guarantee that its content is well-researched, credible, peer-reviewed, or fact-checked. The content feeding the model is the basic problem. It may conduct research on some documents that include inaccuracies. While doing its best, ChatGPT cannot guarantee that the data it uses is the best available. An article that was written and then had to go back and retract some of its content is one possible example. ChatGPT may make references to the article’s main text without taking into consideration the retraction that was afterwards appended. Moreover, ChatGPT is currently limited to information as of 2021. It is unable to react to recent events or accurately reflect circumstances that may have altered after 2021. Last but not least, a word of warning: Any content produced by ChatGPT still needs to be verified, referenced, and double-checked.
2. ChatGPT is NOT human-like in its comprehension
   The ability of ChatGPT to comprehend and write may be frightening those who earn a profession by creating, editing, or grading content; yet, ChatGPT can’t write like a real person just yet. An excellent illustration of its limits may be seen in an article from The Atlantic by Ian Bogost, who asked ChatGPT to assess a poem that ChatGPT had produced. I am a huge language model trained by OpenAI, and I am unable to assess the value of literary works or make aesthetic judgements, according to the ChatGPT response. Human idea and opinion may not always be accurate, admirable, or fascinating, but they are always capable of having one. Moreover, ChatGPT struggles with a number of very easy actions that would come naturally to people. Check out this illustration from the list of ChatGPT failures on GitHub:

Question – Robert has two sons, the user. Jay and John. Jay’s father and brother are both. Two sons are the father’s. There is a father and a brother for Jay’s sibling. Who is the brother of Jay?
ChatGPT: It is impossible to identify Jay’s brother based on the facts presented. The data offered is insufficient and conflicting.
It’s not hard to figure out who Jay’s brother is, and no other information is needed. ChatGPT is unable to comprehend that.

Hi, I’m Garima and I write about life experiences. I have several books available on Amazon. Check them out today! Any purchases or KDP reads will be greatly appreciated. If you like my books, do leave a review. Here’s my author page on Amazon – https://www.amazon.com/-/e/B0BQDZXYNV