A few weeks ago, a former colleague sent me the link to a free, online training programme conducted on behalf of ChatGPT (Chat Generative Pre-Trained Transformer). I had a few days earlier read a column by Thomas Friedman of the New York Times in which he expressed his amazement over the giant leap artificial intelligence (AI) had made.

For starters, he is one of the world’s most widely-read columnists. One of his books, The World is Flat, was critically acclaimed for his innovative approach, however flawed his theory that those who eat McDonald burgers won’t go to war with one another. Clans are better fighters than total strangers as the Mahabharata proves beyond a shadow of doubt.

He interviewed the head of an AI company, which, on his prompting, demonstrated the writing skills of an AI-enabled laptop. For instance, Friedman asked the machine to write an article on a particular museum in New York, USA, in 800 words. He asked it to rewrite using only 400 words.

Then he asked the machine to write the same piece in a Shakespearean Sonnet format. All this was done in a matter of seconds, not minutes. It was against this backdrop that I joined the 90-minute training programme from the comforts of my home in New Delhi.

He dazzled the audience by claiming that ChatGPT, since its release in 2022, has been ruling the roost. It got its first million subscribers within a month of its release. In terms of popularity, it has overtaken search engines like Google. It can do work which the search engine cannot.

In order to remain in the market place, Google was also compelled to introduce its own version of ChatGPT. I am yet to try it and I am, therefore, unable to comment on it, let alone make a comparison. My knowledge about AI is rudimentary, to say the least.

A young man with considerable experience in computing took the class in which hundreds of people from all over the world participated. What struck me was a spelling error in a word used to explain the prowess of AI. As it was an interactive session, I pointed out the error, which many would have certainly noticed.

I got a fair idea of AI from his speech. I wanted to use it to test and review a book I read, and written by a close friend, Chetna Keer, an exceptionally brilliant humorist, whose command of the English language I admire. I have been with her throughout her literary journey. It was to Chetna that I turned when Deepalaya wanted to bring out a book to celebrate its 40th anniversary. After so much of brainstorming, we titled the book Fruitful Forty.

For the test, I gave inputs in 500 characters on Garnets Under My Gulmohar, her latest book. Immediately, I got the review. It was a book of fiction, a thriller. Since I mentioned that the author was an environmentalist, it presumed that the book was on the environment. In short, the review was totally useless. What’s worse, it gave a chapter-wise review, totally uncalled for.

As I mentioned, my experience and knowledge of AI are limited. The India Today news-channel recently showcased its AI skills, when it introduced its first AI-based newscaster at a prestigious conclave it organised with Prime Minister Narendra Modi on the stage.

The channel claimed that it would soon have the PM interviewed by the artificial newscaster, who can speak a host of languages with equal felicity. It was an editor who alerted me about India Today’s leap of faith in AI. Otherwise, I might have missed it.

I wondered whether Modi would subject himself to such an interview. I have seen an interview he had given to Karan Thapar while he was the Chief Minister of Gujarat. He had to ask for a glass of water, thank him and walk out of the interview because Thapar asked him an inconvenient question. Since then he has not given an interview to any journalist worth his name, though he allowed film star Akshay Kumar to interview him.

Since the AI interviewer can be programmed by the Prime Minister’s Office, Modi may find such an interview more comfortable. Let’s not jump the gun. Let’s wait for Sana, the Aaj Tak AI newscaster, to interview the Prime Minister.

After watching the India Today programme, I met and heard the first Malayalam AI news-reader whose video was uploaded on YouTube. The “girl” resembled a film star I knew but could not recognise. After her introductory remarks in beautiful Malayalam, the creator of the video came on the screen.

He explained in simple language how he created the video using easily available resources in the public domain. As I write this, I can also create an AI newscaster who would resemble, say Hema Malini or Amitabh Bachchan.

I can also create text and the newscaster can read it in Malayalam or Marathi, with fluency that I cannot. What I meant to say is that the AI technology is available at virtually no cost to those who are interested. Let me be frank, the video in Malayalam helped to demystify AI and make it intelligible to the masses.

Nonetheless, the threat AI poses is real and sanguinary, not imaginary. It forced thousands of artistes and technicians in Hollywood to go on a strike against the increasing use of AI in film-making. For instance, the producer can easily create lyrics, prepare a screenplay and make shooting schedules using AI. Who knows he can also dupe actors using AI technology.

This is not a figment of the imagination. Someone in Kerala produced a video using a character who resembles thespian Mammootti when he was younger and his hair was black. The question that arises is: why pay enormous sums of money to actors like Mohanlal and Aishwarya Bachchan when actors who are more handsome and beautiful can be created using AI?

I remember the Marxists in Kerala protesting against the use of computers some 40 years ago. What guided the Hollywood workers in the capital of Capitalism when they struck work was the same fear of losing jobs. There can be no disputing the fact that a lot of jobs can be done by AI. For instance, a condensed version of a book can be obtained in a matter of seconds, if AI is given access to the book.

Artificial intelligence is not yet perfect. It can even write and write without facts and figures. For instance, when I asked ChatGPT to write a piece on me as “a senior journalist”, it produced one in a few seconds. 

It was a long piece on the attributes of a good journalist with no relation to me. There was not a single sentence from which I could make out that it was about me.

The article was like the intelligent child who wrote about the cow when she was asked to write about the tree. Since she had prepared herself to write about the cow, what she did was to tie the cow to a tree and describe the cow. The child should have been given marks for her critical thinking, if not her essay on the tree.

In the short film made using a variation of Mammootti, there was one thing distinguishable. If you see the close-up of the face, you can realise that the skin tone and colour are the same everywhere. Ask any photographer, he will tell you that however great the Photoshop’s capabilities are, it cannot indistinguishably fix the head of a person to the body of another. Ganapathy with the head of an elephant is in the realm of faith, not AI.

As of now, AI cannot be a substitute for human intelligence. It can even land a person in trouble. If a boss starts using flowery language in his writing, it will not be difficult for his secretary to identify the AI source. Ditto for the secretary, whose extraordinary linguistic talent flourishes all of a sudden. In the case of AI, what's good for the goose is good for the gander too!

The other day, I heard a speech in English delivered by a person whose command of the language was nothing commendable. When he read out his speech in grammatically correct, idiomatic  language, I could make out that he either plagiarised it or used AI to create it. Of course, I admired his understanding of AI and his willingness to use it in his daily job of preaching.

As the experts learn and improve AI, a time will come when AI will produce books like Shakespeare’s As You Like It and create lyrics like Mah Nà Mah Nà… which means nothing and hymns like Amazing Grace, banned from India’s Beating Retreat programme. Allowing AI to rule the world is also dangerous. 

In Kerala, AI-assisted CCTV cameras were installed by the traffic police all over the state. The camera caught one Orthodox priest for not wearing the seatbelt. He was fined. He refused to pay the fine as he was wearing the seatbelt at that time. The villain of the piece was his flowing jet-black beard in which the belt merged.

When the police officers were convinced, they cleared him of the charge. A few days later, he was caught again by another camera at another place. For him, the question was Shakespearean: To have a Beard or not to have a Beard.

At the rate at which AI has been growing, there is a fear that it might devour its creators, as often happens in science fiction. Imagine, AI is able to guess what a person thinks. There was a bizarre case in which the husband was punished for “killing” his wife. He had in a moment of passion said, “I want to kill you”. He said this in love. He did not know that he was providing clinching “evidence” of his desire to kill her!

If, suppose, in a moment of anger, a person feels like “killing” another person and AI is able to detect the goings-on in his mind, he can be punished for “planning to commit murder”. The dangers of such a scenario are too many to recount here. The European Union has woken up to the threat AI poses and it wants laws to be in place to control the Behemoth that AI has already become. Countries like the US and India will, sooner than later, be compelled to think on the EU lines.

AI will take a few more years, if not decades, to read the human mind. As of now, it is just a tool in the hands of human beings. It requires intelligence to use AI effectively and efficiently. 

Otherwise, it will end up like when a lazy sub-editor of the now-defunct Observer of Politics and Business, a newspaper Dhirubhai Ambani launched in the nineties, used automatic spell-check and published the front page where Deputy Prime Minister Devi Lal’s name was published as Mr Devil consistently in the text.

Books have begun to be written using AI and film producers are increasingly turning to AI to reduce the cost of film production. Today, a young girl can produce a movie with just a mobile phone in her hand. In fact, a few years ago, my neighbour Anil Kumar’s teenaged son produced a short film using a smartphone.

Could anyone have imagined this twenty-five years ago when uploading a three-minute video, produced in India, from Russia or Hong Kong cost tens of thousands of rupees? Artificial intelligence may have greater strength than human intelligence but it will never be a substitute. After all, the created can never replace the creator! 

ajphilip@gmail.com