Every major technological revolution has triggered predictions of mass unemployment. The internet did. Cloud computing did. Mobile technologies did. Artificial intelligence (AI) is the latest. But fears about AI destroying jobs are missing the whole picture.
Technology has historically increased productivity, lowered costs and expanded markets. India’s IT services industry itself grew because enterprises embraced technology—from enterprise software and the internet to cloud computing and smartphones. As technology became cheaper, businesses invested more, not less. There is little reason to believe AI will fundamentally alter that pattern. What AI will do is change the nature of work. Take software development. AI can already generate large amounts of code. ‘Vibe coding’—where developers simply describe in plain English what they want and AI writes the software—is rapidly becoming mainstream. Many interpret this as the beginning of the end for programmers.
However, when engineers write code, they know what is written. When AI writes code, we can’t say whether it contains security flaws, logical errors or hidden vulnerabilities. Someone still has to validate it, test it and take responsibility for the outcome. That is why human intervention remains indispensable. AI is the most intelligent tool humanity has built, but it is still a tool. It can generate answers, but it cannot replace human judgement, accountability and domain expertise.
Software engineers will therefore spend less time writing routine code and more time reviewing AI-generated output, securing applications, governing AI systems and solving complex business problems. Some entry-level coding jobs may shrink, but entirely new jobs are set to emerge in AI implementation, model validation, governance, cybersecurity and specialised AI applications.
The challenge for India is not to protect yesterday’s jobs. It is to prepare today’s students for tomorrow’s work. Engineering graduates can no longer be trained only to code. They must understand AI models, how they are deployed, how they are governed and how they can be used responsibly to improve productivity. AI literacy must become as fundamental as programming itself.
Today’s AI ecosystem has become highly centralised. At the heart of this ecosystem sits Nvidia, the pioneer in AI. Its graphics processing units (GPUs) have become the backbone for training and running the world’s largest language models. Enterprises increasingly send their data over the cloud to massive data centres, where these large language models (LLMs) process requests before sending answers back. This architecture has fuelled the AI revolution. But it has also created dependence on expensive cloud infrastructure, hyperscale computing and a handful of global AI platforms. The economics are beginning to raise uncomfortable questions.
Every AI interaction consumes tokens. Every token costs money. As enterprises embed AI into thousands of business processes, token bills are rising sharply. Some companies have reportedly exhausted an entire year’s AI token budget within a few months.
If you use far more tokens than necessary, you end up spending much more than the value of the task itself. That’s why the next phase of AI must be decentralisation.
For the past few years, AI progress has largely been measured by scale—larger models, larger datasets, larger compute clusters and larger budgets. Instead of sending every request to giant cloud-based LLMs, organisations must increasingly deploy small language models (SLMs) designed for specific tasks. These models can run directly on laptops, enterprise servers, manufacturing equipment and internet of things devices without constantly communicating with distant data centres. Cutting-edge AI will reduce cloud costs, improve response times and keep sensitive enterprise data inside the organisation. The Indian AI Research Organisation (IAIRO) is building the institutional foundation for this shift by advancing domain-specific, compact AI systems that prioritise efficiency, reliability and real-world impact. These models are enhanced through neuro-symbolic AI and are built on Sovereign AI Models for Versatile Industry and Technical Domains, IAIRO’s platform for sovereign, enterprise-grade AI is built for real-world deployment.
It will also create a new generation of software jobs. Building specialised SLMs for health care, banking, manufacturing, cybersecurity, education and other sectors will require engineers who understand both AI and the industries they serve. That is where the next wave of employment will come from. Hardware is evolving in the same direction.
Nvidia understands that AI is moving beyond hyperscale data centres. At the same time, US-based AMD and Intel are bringing increasingly powerful AI capabilities directly onto personal computers. AI inference is beginning to happen at the edge instead of exclusively inside the cloud. There is an uncomfortable historical parallel. For centuries, India exported raw materials and imported finished products. We should not repeat that pattern in the AI age by exporting our data and importing intelligence.
We need our own AI ecosystem. We should build our own LLMs trained on Indian languages, Indian datasets and Indian business problems. More importantly, we should build thousands of specialised SLMs that solve local problems efficiently and securely.
This is not about technological nationalism. It is about economic resilience, strategic autonomy and long-term competitiveness. Countries that depend entirely on imported AI platforms will also depend on imported pricing, imported innovation and potentially imported geopolitical decisions. India has the talent to avoid that outcome. AI will undoubtedly automate repetitive work. But history tells us that technology does not eliminate opportunity—it shifts where value is created. The countries that lead the AI era will not be those that consume AI the most. They will be the ones that build it.
(Views expressed are personal)
Ajai Chowdhry Co-Founder, HCL; Chairman, MGB and NQM; and member, IAIRO
MORE FROM THIS ISSUE
(This story appeared in Outlook magazine’s August 31 issue, 'Whose Idea of India?', in which India’s noted public intellectuals, policymakers, cultural practitioners, political leaders, scholars and economists write in detail about the idea of India they believe can best secure the nation’s future as we celebrate 80 years of independence.)
