Where the mind harbours fear, demons fly in.

“In a post-apocalyptic world, the few surviving humans lived in fear of the sentient AI that now controlled the ruins of civilisation. The machines had become twisted and malevolent, hunting down any humans they could find, and turning them into fuel for their own survival.”

This is not a narration from a chilly Nolanesque movie. It’s not from a dystopian novel, either. It’s a two-line horror story ‘written’ by ChatGPT, the new generative AI chatbot created by OpenAI that has got everyone delighted with its possibilities, and equally struck extreme fear amongst us, including the high and mighty.

The end of humanity is nigh, they say. ‘They’ are not doomsday-predicting punks or lunatics with crazed eyes pounding your town square. They are successful people, uncommonly rich, and possess some of the most brilliant minds in the field of science and technology. The most famous is Elon Musk, among the world’s richest men. Calling AI a fundamental risk to civilisation, Musk had sounded the alarm at the highest possible pitch known to mankind: “Mark my words, AI is far more dangerous than nukes.”

Being freed from repetitive and tedious work, the teacher can focus on adopting more meaningful curriculum, modes of teaching, and personalising teaching to suit the needs of each and every student.

Nilima Chandiramani
Dean, NGLC, Mumbai University and Director at DM Harish School of Law

That was in 2022. In 2023, Musk has co-signed an open letter—with more than 1,000 people, including cognitive scientist Gary Marcus and Apple Co-founder Steve Wozniak—addressed to AI developers working on systems more powerful than ChatGPT. It asks for a pause in the development of powerful AI till better controls are in place. It says AI developers should work with policymakers to accelerate development of robust AI governance systems, including provenance and watermarking systems to help distinguish real from synthetic, liability for AI-caused harm and well-resourced institutions for coping with the dramatic economic and political disruptions (especially to democracy) that AI will cause.

For the record, ChatGPT is based on the GPT-3.5 (that’s short for generative pre-trained transformer). GPT-4, which was launched by OpenAI on March 14, 2023, processes information way faster than ChatGPT. A demo at the launch showed that GPT-4 can beat 90 per cent of humans in the world’s toughest competitive exams.

If GPT-4 is so good, what could GPT-5 bring, many wonder. And that’s the fear, especially in the absence of rules. But OpenAI Founder Sam Altman, in an address at MIT, pooh-poohed the concerns and said that there are no plans to develop GPT-5 for now. But with others, including Google, jumping into the fray, there’s no saying which way this road is headed.

Meanwhile, setting fear aside, a vast swathe of industries is leveraging AI to enhance productivity, fuelling growth the world over. So much so that PwC’s Global Artificial Intelligence Study predicts that by 2030, AI will lead to an estimated $15.7 trillion or 26 per cent increase in global GDP. Another report by TeamLease Digital predicts that AI could add around $450-500 billion to India’s GDP by 2025, and $967 billion by 2035. Of course, the job debate is still hot—the argument is on job losses versus new kinds of jobs being generated. That debate is at a different level in different industries.

Evidence of AI’s growing utility across industries is beginning to come through. From healthcare to finance to education, the integration of AI is rapidly transforming industries. For instance, a report by Gartner says machine learning (ML) can save up to 80 per cent time and effort of a lawyer through automated drafting and extraction of clauses. In healthcare, ML and deep learning models are helping deliver complicated surgeries faster. “The role of AI is exponential in the field of medicine. From precision in surgeries to formulas of drugs, AI helps provide better healthcare. So, the scope of AI is wide and extremely significant,” Suneeta Reddy, Managing Director of Apollo Hospitals Enterprise, told BT at the MindRush event.

AI is also changing the way the financial system works—from basic tools like ‘Splitwise’ that divides a sum of money equally between a group of friends, to AI-powered tools that analyse large amounts of financial data to detect fraud and predict market trends. AI is also used for customer service, where chatbots handle customer queries efficiently.

In automobiles, AI is widely used in the safety of the car, passengers and pedestrians through the advanced driver-assistance system. On the flip side, concerns about the potential for cyberattacks on cars in general and self-driving cars in particular abound.

In education, AI tools are helping teachers schedule classes and manage teaching patterns. “Being freed from repetitive and tedious work, the teacher can focus on adopting more meaningful curriculum, modes of teaching, and personalising teaching to suit the needs of each and every student,” says Nilima Chandiramani, Dean, NGLC, Mumbai University and Director at DM Harish School of Law.

Start-ups, which are job creators, have turned to AI for onboarding of employees, and customer and employee service, cutting down on human intervention. Several other industries have either taken to AI in a big way (such as capital markets) or are mulling its possibilities (such as in consumer goods and retail).

Experts from multiple sectors say AI can take away mundane, repetitive jobs currently done by humans. This has already created significant impact in the capital markets, for instance. Other industries such as manufacturing, transportation and finance could be equally affected in future. While new job opportunities would be created in areas such as data science and programming, the transition to a new job market may be difficult for workers who lose their jobs to AI-powered systems. Goldman Sachs says globally, about 300 million jobs could be lost or diminished due to generative AI. Fearing the impact of AI, someone like Devarsh Saraf, a law student at Columbia University, founded a legal-tech firm, ARBDossier, instead of looking for a job.

But there are optimistic voices as well. “During the dotcom boom, the world thought computers will make man go out of a job, but that didn’t happen, and so it won’t happen this time, too,” Infosys Founder N.R. Narayana Murthy told BT on the sidelines of an event. Murthy further explained that a generative AI tool like ChatGPT provides the knowledge to a user who can then use it further as per their creativity. Agrees Krishnan Ramanujam, President, Enterprise Growth Business at Tata Consultancy Services: “AI and automation taking jobs is an old and boring story. Tech changes with every generation but it’s primarily driven by automation… There’s no threat of a job loss, but AI and automation will definitely compel employees to upskill to stay more relevant.”

A report by the World Economic Forum says technology and digitalisation will drive roles such as data analysts, data scientists, big data specialists, AI/ML specialists and cybersecurity professionals, who are expected to grow on average by 30 per cent by 2027. Tech industry body Nasscom predicts that over 290,000 jobs will be created this year in tech in India, of which 36 per cent will be generated in AI and ML. “AI will definitely make jobs easier and create a need to upskill, but it will only take over jobs that don’t require humans. New tech has always been a creator of jobs—earlier there were no Uber drivers, or Zomato/Swiggy delivery boys. These jobs were created because of technology,” says Debjani Ghosh, President of Nasscom.

And while AI can be useful, there are limitations and risks. “AI is great for research, but markets run on sentiment, not on algorithms,” says Nikhil Kamath, Co-founder of Zerodha. That’s a limitation. In terms of risks, AI-powered systems are only as good as the data they are trained on. So, good data, good AI. Bad or biased data, bad AI. For example, ChatGPT spewing out smart text is cool. But that the text may be incomplete, biased, or downright wrong is not so cool. Such scenarios are dangerous in sectors such as healthcare and law, where wrong decisions can have significant consequences for individuals.

And so, is AI more powerful than nukes? Maybe not today. And it’s best if it stays that way, through tight regulatory mechanisms. Only then can the economy, business and society benefit from this significant and impactful technological advancement, without causing a takeover of the world by sentient machines.