Technology

India’s AI moment: Balancing global technology and self-reliance

The dot-com crash mostly hurt investors. An AI correction, built on borrowed money, could hurt a lot more people

Rajesh Seth

Artificial Intelligence is being increasingly adopted and used in India by both individuals and businesses to manage their day-to-day operations. There has been an interesting recent development in the AI space in India which can benefit Indian businesses. OpenAI, the company behind the popular ChatGPT, has announced that eligible Indian businesses and developers can now have certain requests processed within India.  This announcement raises an important question: can India benefit from the world’s most advanced AI technologies while simultaneously building its own capability to develop and control AI?

Why process in India?

Until now, an Indian company could store its information in India but still have to send it overseas for processing by an AI model. Imagine keeping your jewellery in a bank locker in Delhi but having to send it abroad if you need it assessed or catalogued. While it is away, you have less control over who handles it and how securely it is treated. Something similar can happen with digital information: it may be stored in India but sent overseas for AI processing.

For businesses handling sensitive information, this could be a significant concern. For instance, banks, insurance companies and healthcare providers deal with confidential customer and patient data. Regulatory requirements, internal policies and customer expectations may restrict where such information can be processed.

OpenAI’s recent announcement addresses part of this concern. For supported applications using GPT-6 Luna (a software version of OpenAI which can handle large volume of tasks economically and quickly), both the processing of requests and the computing that generates the AI response will take place in India. The facility is intended for organisations using OpenAI’s business technology services, rather than ordinary users of the ChatGPT application.

Who stands to benefit?

Consider an Indian bank processing thousands of loan applications. AI could help examine documents, summarise financial information and assist employees in identifying incomplete applications. A bank that was hesitant to use overseas processing for sensitive information may now have another option.

A hospital chain could use AI to summarise medical reports, organise administrative records and assist with routine documentation, subject to appropriate safeguards. An insurance company could use it to examine claims and policy documents, helping employees handle large volumes of paperwork. Manufacturers could analyse production reports, quality records and supply-chain information more efficiently.

The opportunity is not restricted to large corporations. Indian start-ups and smaller technology companies could also build specialised applications using OpenAI’s models without having to develop a powerful AI model from scratch.

India’s own AI capabilities

The Indian Government has also been focusing on developing indigenous AI models. On November 3, IndiaAI Mission released three indigenous foundation AI models EKA -1,----2B and -5B. EKA-1 is a ‘server room’ model tailored for the BFSI sector and the Dhir duo models are more portable, suitable for smaller applications of universities, mid-size businesses, small Government offices etc. These models are designed for applications including Indian-language communication and specialised business needs and can operate within an organisation’s own computing facility.

The Government is also encouraging the use of Indian-developed AI systems in strategic sectors such as defence and aerospace, where dependence on foreign technology has concerns of security and control.

Infrastructure behind the ambition

Developing and running advanced AI requires enormous computing power, specialised computer chips, reliable electricity and sophisticated data centres. The Tata Group (via Tata Consultancy Services and its AI infrastructure platform HyperVault) has forged a strategic alliance with OpenAI to build high-density AI data centers to scale AI across India. The data centre capacity will start at 100 MW and scaled upto 1 GW.

Balance, not isolation

India should neither reject foreign AI technology in pursuit of self-reliance nor should it become excessively dependent on a handful of global technology companies.

Foreign AI models can provide Indian businesses with access to sophisticated capabilities. Indigenous models can gradually strengthen domestic expertise, support Indian languages and address applications where security, local knowledge or control over technology are important.

Ultimately, India’s AI success should be measured not simply by the number of models developed or data centres built. It should be measured by how effectively AI improves productivity, strengthens public services, supports businesses and creates opportunities for its people.

The opportunity before India is to bring the world’s AI capabilities to its businesses while building the skills, infrastructure and home-grown technology needed to shape its own AI future. The real ambition should be not just to use AI in India, but to make India an important contributor to the global AI ecosystem.

The author is an alumnus of IIM Ahmedabad and Professor of Practice at IILM University, Gurugram with an interest in AI, Technology and Strategy; Views presented are personal.