Marketing Intelligence: October Issue

By: Aaditya Jain

Ever used a chatbot on a website? How often do you feel satisfied with its response? Well, most of the times, one’s question remains unanswered, and then, one is compelled to pick up the phone, and dial customer care. This is because of the extremely limited database on which the chatbot has been developed. It is not stored with most of the answers to the customers’ questions.

A phenomenon that has recently come into limelight (as early as Nov 30, 2022) is that of ChatGPT, which stands for Chat Generative Pre-trained Transformer. It is a much advanced version of chatbot. ChatGPT is developed on an extremely huge database, and holds the ability to make decisions based on that database. This database includes emotions as well. Yes, ChatGPT has the ability to judge the sentiments of the person by virtue of the tone and punctuations, and respond accordingly. In other words, ChatGPT uses Artificial Intelligence (AI) in addition to Machine Learning (ML), as opposed to chatbots, which are limited to only ML.

Moreover, one of the beneficial features that it offers is “AI classifier”, wherein, ChatGPT marks its responses with this identification, letting the users know that the response is subjective, and not entirely based on facts. This feature helps the users in judging the authenticity of the response.

When you face issue with one of your home appliances, and call a technician, don’t you feel that you could have repaired the appliance yourself had you been guided over the phone? There wasn’t any need for you to wait at home for the technician to come? ChatGPT can be used by companies to provide the customers with such plausible solutions, which would prove out to be way more effective than the current chatbots, and even the customer care executives who make you wait over the phone for a whole lifetime to only let you know, at the end of the call, that a technician would need to visit you.

This would prove out to be beneficial not only to the customers, but also to the companies in many ways. Marketers who want to write a promotional offer, but don’t know where to start, or how to structure their thoughts, can just ask ChatGPT to do it for them. The initial draft would be ready in no time. But, proof-reading and fine-tuning the piece is required commensurate with the product’s exact specifications and the target audience. Don’t worry, this article has not been written by ChatGPT!

ChatGPT can also boost revenue for companies by providing customers with all the information that they require regarding a product (instead of screening through various articles), making the customers reach the First Moment of Truth a lot earlier. All that the customers need to do is ask, and counter-ask ChatGPT. Leaving aside the one-time initial cost of deploying ChatGPT, companies would save leaps and bounds on their on-the-ground technical and customer care workforce, and office space. Companies would also improve their brand equity by reducing the downtime of their products by virtue of the consumers themselves fixing the issues with just a little guidance from the ChatGPT.

One other interesting application of ChatGPT can be to facilitate interactions between humans and robots. If ChatGPT can be deployed in robots, manoeuvring their actions and behaviour as per the response to the questions asked or instructions given by a user, a lot of tasks would be easily carried out by the robots. However, a caveat here is – in the interest of a greater good, robots should not harm humans under any circumstance.

However, a nightmare - ChatGPT making the “Dead Internet theory” come true. A lot of marketers copy-paste the content provided by ChatGPT for their marketing campaigns. Since the algorithm, basis which ChatGPT gives its opinion or provides content, can be biased towards the agenda and propaganda that is believed in by its developers, so can be ChatGPT itself. This could fill the internet with one-sided information, creating a misbalance, and ultimately, leading to a greater disaster - brainwashing of the youngsters.

One of the limitation of ChatGPT is “Jailbreaking”, wherein the users engineer their questions/prompts in such a way that ChatGPT, uncontrollably, moves out of the restrains that have been set on it, and gives its opinion and provides content on violent and A-rated topics. By the way, it was not easy, and completely justified, to set these restrains. In one of the cases, some Kenyan workers earning less than $2 an hour were called to make the ChatGPT learn from their responses when exposed to extreme abusive content.

Moreover, ChatGPT has very limited information on the events that have occurred post 2021. It has only been fed with the data prior to that. And because its database is not updated in real-time (in contrast with the internet), ChatGPT should not be completely trusted with performing data analysis and formulating business strategies.

Nevertheless, at the end of the day, machines can be used to carry out certain tasks, but, should never be given the authority to take crucial decisions. Decision making must be carried out by someone with emotions, by someone who holds the fear of facing the consequences of those decisions – humans. Even if it means less big data analysis, and less business. This is probably the reason why countries such as China, North Korea, Russia, and Iran have already banned ChatGPT.