AI (Artificial Intelligence) has always been a mystery for most people. We only think of Jarvis in Iron Man and what the movies tell us of sentient androids behaving like humans but way smarter. That perception is still too far into the future as we must take baby steps first.
With the arrival of big data in recent years and further technological advances in areas like Cloud Computing and Machine Learning, it is possible to build an ecosystem entity that is able to absorb unlimited information to provide responses that amaze us.
ChatGPT https://en.wikipedia.org/wiki/ChatGPT is one such recent milestone that has already started a global AI arms war to bring us to the next quantum leap. It was the fastest app to gather a million users within 5 days when it was soft-launched in Nov 2022, a mere 3 months ago.
ChatGPT was a novelty at first, it was fun to ask questions to a new chatbot and to task it to provide a simple and yet very comprehensible answer. Its responses were next-level and totally blew our minds. I started to play with it and was impressed with the answers which sounded more intelligent that the ones we are used to from Google and other chatbots.
We have reached a point where the latest chatbots like ChatGPT have most likely been able to pass the Turing test https://en.wikipedia.org/wiki/Turing_test to make it indistinguishable from chatting with a human. One of my NUS lecturers had been toying with the creation of 2 distinct chatbot profiles from OpenAI and posted the conversations he has had with them on LinkedIn. He just published a book on this experiment called “Bedtime Stories from an AI”. https://www.amazon.co.uk/Bedtime-Stories-AI-intelligence-changing-ebook/dp/B0BHNGXCH2
Soon, many enthusiastic users found new ways to utilize this tool. The more we played with it, the more users discovered its potential to seriously disrupt the way we do things. It may give wrong answers once in a while but generally, the longer responses seem like a well-thought-out writeup which is created in mere nanoseconds.
One can prompt ChatGPT to go into a frame of mind before tasking it to respond to a particular question based on the parameters given. By assigning it as a specific persona right at the start, the user can help it narrow down its search for a well-thought-out reply that is more relevant to the question asked. For example, start with: “Imagine you are an interviewer/CEO/scientist” etc. That sharpens ChatGPT’s response and provides an even more impressive answer that is specific to the needs of the user.
Just a few days ago, I asked it to help me fine-tune my LinkedIn profile. The advice was that my write-up needed to sound more professional and it reworded my intro to become more concise and impactful. A number of tips were also given for me to think about. I even suggested to my mentee to use it to help to tighten the resume cover letter as I felt it was too lengthy and not concise enough.
The sky is the limit now as other similar search engines feel threatened and want to showcase their challengers to ChatGPT. Google rolled out Bard to protect its turf as the most used search engine. Chinese names like Baidu and Alibaba are also doing the same. They do not want their cheese to be totally eaten by ChatGPT even as Microsoft is now integrating it into its Bing web browser.
There are so many more ways yet to be discovered where this new AI technology can be utilized than we can ever imagine. I had the chance to work with a firm to develop a chatbot for a Myanmar microfinance a few years back. It was tedious work trying to train the bot just to answer simple client queries on the company website. A question had to be asked in multiple alternative ways just to ensure that the machine can learn to provide a credible response to the end user.
ChatGPT has already been pre-trained and had absorbed all information up to 2021 from the World Wide Web via sites like Wikipedia and much more. Being already trained on the huge database of info, it can then be twitched and customized to any type of Chatbot requirements, be it business or leisure focus.
With so much recent hype on the Metaverse, it clicked on me that ChatGPT might finally be the missing piece that can create a virtual universe with avatars living through immortality can exist. Let me backtrack this a bit to expand on my hypothesis.
The very good dark humour British TV series Black Mirror had an episode that won many awards https://en.wikipedia.org/wiki/San_Junipero a few years ago. What if one can live forever in a virtual world once the physical body ceased to exist or expires? That was the premise of the plot which the star of the show had to grapple with, to stop existing when one dies or having one’s mind ported to an alternate world to live forever. She found a new love in that virtual world while test-driving and was faced with a hard decision on whether to enter it before her death.
From the many ChatGPT articles I had read so far, it seems that one can train it with one’s personal data to create an artificial “you” that will talk like you. If we combine this with a human-like digital avatar and artificial speech synthesis, we can technically create a digital alter ego that can live in a Metaverse universe. Hence, one can technically sustain one’s “digital self” or even create “digital love ones” of persons who have died. These can then be placed in the new Metaverse and live all the way to immortality with similar avatars in the simulated virtual world as it will be inhabited by similar “intelligent digital entities” of many other like-minded persons.
And where can we get enough personal data to do that, you may ask? Well, you already have a digital device that you cannot live without on your hands. Your mobile has been capturing your behaviours and life history since your first iPhone. Cloud backups ensure that it continues from an older to a newer model. Each of our personal Whatsapp accounts alone already has at least 5 to 10 GB of personal data on how you interact with all your social contacts over many years.
We can now feed all the above data from our mobiles to train ChatGPT to respond like a customized and personalized chatbot of oneself. All this looks scary and spooky but isn’t this feasible now?
This is just one of the many ways that the advent of this AI tech will help us explore new frontiers of uses we could never have imagined were possible just a few years ago. Almost 90+% of big data was only created in the last few years. Machine learning and deep/reinforced learning trains better the more data is given to them. With quantum computing, speed will not be an issue to digest more data to provide even better responses.
Whether we like it or not, Pandora’s box has been opened and we will explore the new frontier with AI evolving constantly.
I
Leave a Reply