Vivek Tripathi wanted to write a haiku to surprise his wife on their wedding anniversary. But he did not have the time or inclination to sit down and ruminate, and let his creative juices do the deed. So the bureaucrat, who works on high-speed rail, turned to his new personal assistant.
No, it was not a secretary or some prolific new recruit. Tripathi just instructed ChatGPT, the new chatbot on the internet, to make a haiku for him.
“It came out pretty accurately!” he said, mighty impressed.
Of course, making Japanese poems of seventeen syllables is not the only thing that Tripathi does with this artificial intelligence (AI) tool that has taken the world by storm in the past few weeks. He uses it to schedule his meetings and draft routine emails to employees that otherwise would have taken up a good chunk of his time.
“I use ChatGPT, as well as many other AI tools, to increase my personal productivity,” said Tripathi. This ranges from asking Dall.E, a web application that can give visual answers the way ChatGPT comes up with text, to tailor a meme to send to his son on his birthday to asking ChatGPT to prepare an algorithm for a dynamic pricing mechanism for trains when booking goes beyond a certain percentage. “I’ve become a keen user of these technologies, now that I have seen what it can do,” he said. “AI is going to make a world of difference.”
That world, in fact, is already here. Talk of generative AI or machine learning (ML), where computer networks use data from billions of sources to come up with predictive human-like options, has been doing the rounds in the past few years. But then, they were thought of as geeky gobbledygook rather than anything that really concerned common people.
Not anymore.
AVATAR
Artificial intelligence burst into Indians’ consciousness this winter when a startup halfway across the globe launched ChatGPT three months ago.
Created by San Francisco-based OpenAI, Chat Generative Pre-Trained Transformer is, to put it simply, a free computer program (there are paid, advanced versions now available, including the next generation large language model GPT-4 which came out a few days ago) that can write human-sounding answers to almost anything you ask―from Mughal history to quantum physics with a literary flair. What about an article detailing Narendra Modi’s Gati Shakti plan and how it will change India’s urban infrastructure? Or an ad copy for a washing machine with a humorous twist? No worries! How about a legal argument for Manish Sisodia’s bail application? It gets down to it, instantly.
No wonder then that Chat GPT’s impact has been viral. While some colleges in Bengaluru banned students from using it, worried that they may make it write their assignments, some politicians alleged bias in the results it generated. Fears also abound on it killing jobs in many sectors.
A BRAVE NEW WORLD
However, for the time being, more intriguing has been its across-the-board adoption. From corporate executives in Mumbai using it to simplify office tasks and software engineers in Bengaluru fine-tuning their codes with it to copy writers in Gurugram generating client pitches using it, everyone is having a go at ChatGPT.
Intrepid startups are piggybacking on it to develop unique customer care solutions. Last month, fintech company Velocity came up with ‘Lexi’, which, it calls, India’s first AI-powered chatbot. Wrangler, the denim company, incorporated a GPT-based bot on its social media handles for customers last month. “The bot engaged with customers to celebrate ‘Anti-Valentine’s Week’ , giving funny tips to help them break out of unhappy relationships,” said Sidharth Bhansali, founder of Noesis.Tech, which did the back-end work. “The GPT-based approach allows for a more natural conversation between a customer and an automated system. Unlike the existing models of customer care, it is not limited to a pre-defined list of options. We are excited to see where this technology will take us in the future.”
So ‘human’ are ChatGPT’s capabilities that existing chatbots like Alexa and Siri now actually sound neanderthal. Recently in Hyderabad, Microsoft chief executive officer Satya Nadella (Microsoft is an investor in OpenAI) argued with the bot on stage on whether biryani was a ‘tiffin’ meal or not, with ChatGPT finally agreeing that it was wrong in calling biryani a ‘tiffin’.
AI FOR INDIA
While ChatGPT is the flavour of the season, it is just the tip of the AI iceberg that is about to cross paths with humankind. And India has bagged a starboard side seat for the ride―as per a study by the Brookings Institution, India is only behind the US and China in AI research, and features at No. 8 on the list of the world’s top nations in AI advancements and funding. India has filed more than 5,000 AI patents so far, and at least 10 AI startups from the country entered the unicorn club (companies with a valuation of more than $1 billion) in 2021 alone.
The future could well belong to the neural machines. A report by Accenture predicts that use of AI could raise India’s growth rate by 15 per cent by 2035. The government, of course, has been an active cheerleader. NITI Aayog laid out its National Strategy for AI way back in 2018, and the last Union budget provisioned for Centres of Excellence for AI and robotics at top institutes.
Binesh Sreedharan/ Midjourney
“Make AI in India and Make AI work for India,” said Finance Minister Nirmala Sitharaman in her budget speech. In a post-budget speech on February 28, Prime Minister Narendra Modi reinforced the focus on AI, when he urged people to identify 10 problems in society that can be solved using AI.
Of course, it is not like AI was not being used earlier. From internet recommendations on video and shopping sites to filters used on mobile phone pics, elements of AI have long been in play even on the consumer-facing side. But the technology was essentially hidden within layers of back-end infra.
With applications like ChatGPT and Dall.E, it is now directly in the hands of individuals and small businesses. As Debjani Ghosh, president of the software industry association NASSCOM, wrote in the industry report ‘AI Gamechangers’, “AI no longer remains confined within the walls of the research labs; it’s out of the cocoon and making its impact felt across industries.”
The buzz this ‘AI for the masses’ campaign has created is hard to ignore. References to AI, ChatGPT and related terms have gone up not just on social media and search engine trends, but even in recent corporate earnings calls. Nvidia, which makes chips for complex AI computing tasks, is one of the best performing global stocks this year. The reason could simply be the far-reaching transformation AI promises.
“AI will disrupt all facets of life, going forward,” said Soma Dhavala, director, machine learning at Wadhwani AI, a Mumbai-based institute that studies the impact of artificial intelligence in India and deploys AI solutions for social change. “Any space that is AI-ready will gain from AI. Jobs will change as well, as will many aspects of how we interact with and perceive the world.”
AI is already in use across industries, ranging from the factory floor to customer-facing services. “[It] is being used in almost every sector in India,” said Soharab Hossain Shaikh, assistant professor at the department of computer science and engineering, BML Munjal University, Haryana. Maruti Suzuki, India’s largest carmaker, uses it to run chat bots for its Nexa and Arena sales and after sales services. “These new-age digital tools help us to quickly respond to common queries and provide an enhanced customer experience,” said senior executive officer Shashank Srivastava.
Many companies are capitalising on its seemingly unlimited talent possibilities for anything from cutting costs on the factory floor and implementing new healthcare paradigms to empowering breakthroughs in agriculture and education. The startup Carnot, funded by the Mahindras, runs a tractor subscription service in central India in which the vehicles are hooked up with AI sensors that can monitor crop growth and soil quality.
In fact, what has excited many in India is how AI can be a powerful change agent by bridging the digital divide, by providing the benefits of the internet to those who do not know English or even read or write. “With advances in speech, language and vision processing and understanding, how we interact with technology will become a lot more human-centric,” said Dhavala. “As a result, all technologies, services and products can be made more accessible and inclusive, which presents enormous and diverse opportunities.”
EX MACHINA
“AI has come a long way, and its continued evolution is likely to shape the way we live and work in the years to come,” said Shaikh of BMU. That brings with it its own worries. While job losses are very much a concern since the advent of ChatGPT, bigger worries of privacy and Orwellian scenarios in a world run by machines and data have resurfaced with gusto.
“AI technology can potentially be used for malicious purposes, such as fraud, data theft, and identity theft. It is still relatively new and lacks regulation in many areas, leaving it susceptible to misuse,” said Satya Muley, Supreme Court lawyer.
In fact, one of the initial worries ChatGPT spawned was, in a sign of the times, about the neutrality of its data, rather than the authenticity of it. OpIndia, a right-wing publication, was one of the first to point it out, when it revealed how the chatbot came up with a tirade when asked to write about Narendra Modi.
“ChatGPT carries inherent risks such as hallucinations, biases and the black box nature of its response. The legal implications of the content it produces are not fully understood yet,” said Arun Chandrasekaran, distinguished vice president and analyst at Gartner.
Privacy is another issue. If the algorithms and data analytics of Big Tech in the past decade were consternating enough, AI takes it a step further, considering the amount of data it uses. There are three worry points here: the information the user inputs while asking a question, the data that the machine uses to come up with a reply, and the way in which the resultant answer is used.
“A lawyer may share details of a divorce agreement or a programmer might ask it to check a code. These pieces of information, along with the generated essays, become part of ChatGPT’s database and can be used to further train the tool and be included in answers to other users’ prompts,” said Muley. There is more. “It may happen that the common public might misinterpret the information.”
And that is just ChatGPT. AI systems are already at an advanced stage when it comes to public surveillance using facial and gait recognition. “There is also the looming threat of lethal autonomous weapons like drones and swarms,” said Shaikh of BMU. “AI-enabled terrorism, and lack of explainability” could be a real fear in the future, he said.
JUDGMENT DAY
Chandrasekaran of Gartner believes governments will now step in. “The governments so far have not legislated generative AI but that will soon change owing to deep-rooted implications of these applications and their impact on privacy, bias and other societal issues,” he said.
On the job front, the predictions are mixed―while some fear many jobs could get extinct, others believe AI and ML could spawn a whole new set of skills focusing on managing the new scenario. “Be it ChatGPT or AI, it will never replace humans,” said Infosys cofounder N.R. Narayana Murthy a few weeks ago in Mumbai, pointing out how similar fears abounded when computers came to India, but were misplaced.
The biggest question, of course, is what dystopian nightmares are made of―will AI replace humans, or will it make Sapiens its slaves?
“Ultimately, the objective of AI is to attain human-level intelligence,” said Shaikh. For example, a recent breakthrough at an Aussie lab suggests that ‘brain organoids’, or self-assembling aggregates of neutrons generated in a petri dish using stem cells, can go into creating the next generation of bio computers which could someday, or so some believe, develop a form of consciousness. Whatever the future brings, it is clear that we need to exercise prudence and prioritise ethical practices during the development of AI systems.
Tripathi, perhaps, sums up the ‘brave new world’ of AI best: “We might mess it up rather than AI messing us up!”