Recently, I had talked about OpenAI’s ChatGPT – an AI Chatbot that could explain Code, thermodynamics and writes even cinematic scripts on its own.
The AI is built by OpenAI, company co-founded by Richest of the world Elon Musk, which prospects AI’s capabilities in number of ways. The firm’s latest released prototype ‘ChatGPT’ is nothing but an AI bot that can converse on any topic with users, showing off new capabilities. ChatGPT is available for everyone in its website. It had already achieved 1 million sign-ups to try the AI chatbot, in less than a week.
To have a chat with AI, I too signed up in the site and I asked few questions to the machine-AI and astoundingly the results are clever and methodical, but sometimes iterating and illusive.
Let me show you what I had questioned it and what it had replied.
Chatting with AI – ChatGPT
I started with normal essay-like question about ‘climate-change’ and the answer it had given was conversationally declarative. It seems like it had searched in google-search at the back-end.
As the answer for climate-change was kind of informative, I switched for opinionated questions.
On searching the exact words of the AI on google, there are no links to conclude that AI had literally copy-pasted from Google but it is obvious that it has a library from which it extracts the key-points and constructs them into a sentence.
The AI has been giving out answers whatever I asked, so I asked if he is a genius. It became conversational.
To get it spicier and more controversial, “Whom do you consider the greatest actor in India?”, I grilled.
The answer I got was cunningly wiser by starting with “It is difficult to say” to not mention anyone blindly, but then cited top actors in India like Amitabh Bachchan, Shah Rukh Khan, and Rajinikanth.
While asking the greatest actor in the World, the lines were exactly for what I had queried for that in India, replacing the top-actors in India with that of in the world.
AI might be Deceptive
The AI has specific formats for different question types – for example, if we ask suggestions / advices / examples, it points out one by one. If question is argumentative or opinionated, it tries to back-off by adding some formal-lines and while seeking informative answers, it accents declaratively. In short, it tries to be conversational like human, but could not be fully humanized (at least any time soon). Hence, the AI cannot be considered for now as a source of factual info-seeking platform but just a chit-chat friend with a better knowledge. And most importantly, one should not be liable to the words of AI as it could be misleading.
The most exciting aspect of the AI was that it could able to compare a given topic with the user-mentioned situation and explain that within seconds. For example, I asked it to explain quantum computing like Jack Sparrow. This is what I got. Kind of brilliance.
Finally, I asked for dialogues of Rajnikanth, an Indian actor from TamilNadu, expecting the movie dialogues of the actor. And it started to blabber some ‘superficial comments of oneself’ adding “I am Rajnikanth” at the end. It was fun.
If you need to try the AI and experiment out of it, here is the link to sign-up. You’ll be asked to verify with your phone-number or WhatsApp-number for using the platform.
What you think of this AI-Chatbot?
(For more such interesting technology and innovative detailing, keep reading The Inner Detail).