In recent weeks, two seemingly unrelated developments in the field of artificial intelligence (AI) have brought India into focus. The Global Partnership for Artificial Intelligence (GPAI), an international initiative that focuses on responsible AI use, passed its chairmanship to India. This coincides with India’s presidency of the G20. In parallel, OpenAI, an AI research initiative in the US, made public its ChatGPT system, which has raised the chatbot bar with its ability to write and provide answers in English to questions across a wide range of subjects.
ChatGPT’s interactive interface that was made publicly accessible for testing and feedback has since gone viral on social media, with questions ranging from math problems to software programming source codes being generated by the AI tool. While ChatGPT has certain restrictions imposed on it, the viral social experiment underway through its interface raises several questions on how AI can potentially blur the lines between creative products resulting from human cognition and programmatically generated text that is almost human-like. The ease with which the AI tool is able to generate fiction in response to queries also poses questions on how to assess the truth quotient of artificially generated creative products. As an instance, when asked to compose an essay on events that may not have occurred historically, the AI tool got its dates mixed up. Fictional products from generative AI are perhaps less of a worry than students using it to draft project reports, solutions to assignments, and more.
While regulatory concerns around AI are yet to reach a critical threshold, substantial groundwork seems to have been put in place with the recently released Data Protection draft bill and from early indications around the proposed Digital India Act. A key area of concern would be the use of large personal datasets and AI-based algorithms for the profiling and behavioural targeting of individuals. Such behavioural targeting is already an everyday phenomenon for most users of smartphone-based apps and platforms such as YouTube and Facebook. While the aiming of advertisements based on browsing history is common place and overt, it is the more subtle recommendation algorithms that need scrutiny. Given the propensity of these algorithms to suggest similar content based on past browsing and consumption behaviour, the risks of algorithms reinforcing biases, prejudices and amplifying fake news need to be assessed. In light of recent experiences in the US with politically oriented content on Facebook and Tik-Tok, the manner in which recommendation algorithms can distort the public discourse needs to be put to greater regulatory scrutiny.
While fears around AI are natural, they must not distract from the immense potential technology holds for India’s development. The initiatives of the government under the Digital India banner through the National Programme on Artificial Intelligence are noteworthy for their focus on starting early and engaging school students.
The Digital India week held back in July was particularly striking; young school students showcased their ideas around potential applications based on AI for India’s development to Prime Minister Narendra Modi. From using AI for soil analysis and crop recommendation to helping identify herbs and leaves of medicinal value, the student projects offered a glimpse of AI’s potential in traditional sectors of the economy that are perhaps the last to embrace cutting-edge technologies. Also on display were apps focused on improving athletic performance, doing yoga and the early detection of diseases. With several studies highlighting the potential of AI to accelerate drug discovery and assist in disease management as well as create new biological products, the spectrum of AI applications with relevance to India is quite broad.
Realizing this vast potential, however, will require further strengthening the country’s enabling ecosystem of data centres, cloud computing services and open stack API-based algorithms. While the past few years have seen India make reasonable progress on this front, we are still a substantial distance away from creating formidable homegrown cloud-based alternatives to Amazon’s AWS or Microsoft’s Azure from both a scale as well as ease of use standpoint. While the Bhashini initiative is a significant effort at creating an open-stack algorithm library for recognition, translating and transcribing Indian Languages as a domestic alternative to Google’s language services, similar open stack efforts are needed in other areas such as visual recognition and generative AI, among others. Lastly, more core research in AI needs to happen in Indian research labs and academic institutions. This would ensure that a significant body of intellectual property in this emerging technology space is created in India. The many tools created by the OpenAI project in the US and the extensive use of recognition technologies by China are important lessons on the strategic need for India to invest as much in core intellectual property as the country does in applications of AI.
With G20 and GPAI leadership coming to India, the world is keenly looking at India’s unique approach to digitalization, one that leverages open stack-based digital platforms for mass use. India’s relative success with digital public goods such as Aadhaar identities, Unified Payments Interface (UPI) transfers and Co-Win vaccination in recent years offers a strong foundation to take AI to traditional sectors of our economy in mission mode.
Shashi Shekhar Vempati is the former CEO of Prasar Bharati.
Milestone Alert!Livemint tops charts as the fastest growing news website in the world 🌏 Click here to know more.