Even as the earlier generation was still trying to understand the expanse of Google, ChatGPT, seems to have fired everyone’s imagination. More than a million users were hooked on to ChatGPT soon after its launch by ‘Open AI’, an Artificial Intelligence company. What is it and what are its implications?
Whereas the early Chatbots like the ELIZA (1966) and PARRY (1972) were more interpretive with no intelligence, ChatGPT (Chat-Generative Pre-Trained Transformer) is a dialogue-based deep learning model with three or more layers of neural network that stimulates human brain while reading or writing. It is capable of understanding natural human language, with the ability to answer follow-up questions, admit its mistakes, challenge incorrect premises, and even reject inappropriate requests.
Bot is a software application (SW) programmed for specific tasks or to simulate human activity and is designed to automate tasks on its own without human intervention. Chatbot is also a SW used to conduct online chat conversations via text or text-to-speech, without a direct contact with a live human agent. Automatic speech recognition is the technology that converts spoken words into text and makes Chatbots smart. One can find them in automobiles, home appliances like washing machines, microwave ovens, television remote controls and so on.
Chatbots, are powered by large amounts of data and computing techniques. They have recourse to a vast amount of vocabulary and information, and understand words in context. They then mimic speech patterns and make predictions that come together in a meaningful way. However, its consequences on education, research, employment as well as authoring, is yet unmeasured and unknown which is a serious concern and challenge.
Early this century, the internet changed our educational systems challenging the concept of originality among academia. Anti plagiarism tools were evolved and today “Turnitin” like tools give you an almost 100% check on even a 10000-page document. Interestingly there are also any number of tools available which will rephrase an entire 10000-page document and get you a 0% detection on “Turnitin”. Now the question is no longer about originality, but rather, it is about role, purpose and the learning process.
An oft repeated debate on education is about a great divide that exists in standards of education between the urban rich and the rural poor. Some reasons for the divide are lack of infrastructure, inadequate quality teachers, poor bandwidth, limited access to internet, high cost, and so on and so forth. Despite governments attempts it only seems to widen more, every passing day, with sophisticated technology and more accomplished chatbots taking over everything we do. ChatGPT can democratise education and remove it from the clutches of the elites thereby creating a level playing field for all just as Google has done.
In times to come, the digital avatars of our universities will use ChatGPT like tools causing additional stress to the conventional universities. Teachers will have to adapt from being just teaching to being guides and mentors. Project based learning will replace conventional teaching. However, the challenge for the universities and the regulators will be how to align the future examination system with this technology. Probably, the services sector will improve and in time will eliminate the broker business.
Ask ChatGPT if all that it says is right and out comes the answer. “ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers”. All said and done, the Chatbot will eventually find its way into the classrooms and students note books as it will into hospitals, marketing departments, and everything else, even as its development is Work in progress (WIP).
Even as people are exploring ChatGPT 3, its next version is in the market which can analyse more than just text. At its core, the biggest change to GPT-4 is its ability to work with photos that users upload. Providing more precise responses, it can streamline work across various industries.
ChatGPT is not ‘know-all’. There are several things it cannot do. It cannot predict future results of sports events or political contests and cannot discuss partisan political issues. Further, there are questions it will not answer. Typically, questions that promote hate speech, discrimination, drugs, violence, profanity are not answered. It is of no help if one needs financial or investment advice. That being what it is, let’s make the best as long as it lasts.