ChatGPT has been quite a rage since its launch by the non-profit research company OpenAI in November, 2022. People have been using the conversational Artificial Intelligence (AI)-based chatbot to complete their assignments, writing work e-mails in specific tones, styles and instructions and even to generate creative writing. Many people have admitted to addiction to the new fad. ‘How to use’ guides have appeared overnight, and multiple LinkedIn timelines are full of random memes generated from ChatGPT.
In the words of its creators, “ChatGPT interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.” GPT (Generative Pre-trained Transformer) is a language model that has been ‘trained’ using massive amounts of data from the Internet. It can carry out ‘natural language generation’ or create text responses so astounding that they could have emanated from a human being. This is as close as it gets to having a human behind the screen. AI has been a part of our lives for a long time now, but the conversational mode makes it so appealing, so germane, so relevant and ever so scary.
Began with Deep Blue
As a student in 1996, I watched IBM’s supercomputer Deep Blue challenge Gary Kasparov to a chess match. The match was won by Kasparov. There was a re-match in New York in 1997, in full glare of global media, and this time Kasparov lost to Deep Blue. It was a symbolic moment - Deep Blue’s victory characterized the ‘coming of age’ of machines. By winning against one of the geniuses of the time, Deep Blue was sending a message that artificial intelligence was catching up to human intelligence. While Deep Blue’s win had come from brute computation force which pre-empted millions of chess positions before making a move itself, today’s evolved AI models use billions of data points stored in vast data sets and complex algorithms to generate responses. Each algorithm ‘learns’ and ‘evolves’ itself as data sets become richer. AI will get exponentially more intelligent as the computational and storage capability evolves.
We are all familiar with AI, when we use customer service chatbots, the (almost) ubiquitous Alexa and Siri, auto-sentence completion, grammar correction, etc., on our word processors. However, ChatGPT, which reportedly crawls and learns from millions of webpages and billions of words and data points from the Internet to offer conversation style responses, is a quantum leap.
Full of possibilities
Businesses can improve customer experience and can automate repetitive tasks like responding to customer queries on email and social media channels by using conversational chatbots. Similarly, brand managers will be able to create compelling content for their social media channels, websites, e-mails, blogs and such. ChatGPT can be used to make ad copy that is clear, captivating, attention-grabbing, customized to the target audience and aligning with the brand’s overall messaging. It can be used to create targeted display and email campaigns promoting products to specific customer segments of the brand. Brand and social media influencers can use ChatGPT to generate personalized messages for their followers and build relations with them. With a rich enough data set, ChatGPT can be used to forecast customer behaviour to offer personalized recommendations. It can also be used for sentiment analysis, opinion mining, emotion detection and social listening from customer reviews, blogs and news articles in general. A brand’s website search ranking can be improved by generating optimization elements like meta tags, alt tags, and such, using ChatGPT. Of course, this requires a large and diverse dataset to train the algorithms. In fact, businesses and business managers are limited only by their imagination in the way they can put ChatGPT to use.
Should humans worry?
Doomsday predictions for a lot of companies, businesses and jobs have already begun. These fears may be a bit unfounded. Advancements in transportation and mobility, connectivity, computing, data sciences, artificial intelligence and other fronts will continue. The human race has always evolved and taken any technological advancement in its stride to find suitable applications to make life easy for people. Calculators, computers, robots and more threatened to make humans redundant. But here we are, riding on these very advancements to find newer avenues of staying relevant. It is likely that some jobs will be automated or replaced by artificial intelligence in the future. At the same time, new jobs will be created because of the increased use of AI.
But for the moment, ChatGPT has made my job as a parent a tad easier; albeit until my twin girls discover that the much-appreciated essay I wrote for their school project was copied straight from ChatGPT. Shhhh!
How ChatGPT describes itself
ChatGPT is trained on a large corpus of conversational data, such as movie subtitles, books, and online conversations. This allows the model to learn the patterns and conventions of conversational language, including the use of informal language, contractions, and conversation-specific phrases and idioms.
(*This para was generated on ChatGPT)
(The author is a senior professional in the corporate sector and writes on varied topics that catch his fancy. Views expressed here are his own. He tweets at @sandeepbangia)
