New Delhi, April 16 -- The nuclear research facility, famously known as CERN, also the world's leading particle physics lab, made its World Wide Web code freely available in April 1993. The digital revolution took off. Then came the dotcom age and the infamous dotcom bubble burst, followed by coding and now Artificial Intelligence, or AI. AI is now a part of the common vocabulary like Internet, World Wide Web, digital presence, digital arrest, and hackers. The last two are impacting many adversely, posing a risk of exploitation to the digitally unempowered. Thus, AI is now a Goliath, but with enormous contradictions of hope, greed, and fear. Never before, perhaps, in the world has 'Artificial' impacted so much of the 'Real'.
AI is perceived today as a miracle arriving at the doorstep, yet created by humans. The stock economy is, these days, driven by investment in IT, with AI taking almost the entire cake. NVIDIA became the first company to cross a USD 5 trillion valuation, with Apple breaching the USD 4 trillion mark. There is little chance that AI firms are driven by the spirit of altruism. The quest for more money is visible at the forefront of the AI economy, as we witness Amazon and Microsoft fighting it out over OpenAI.
AI's tremendous potential of arriving at the quickest but informed diagnosis, individual need-matching medicines, precision-enhancing surgical interventions, and many more on the anvil is changing the landscape of medical and health management. Preventing human exposure in potentially life-threatening conditions, and helping in disaster recovery and management are some of the shining examples in AI's ever-inflating bag of achievements. The six questions of the world - which, where, when, why, what, and how - are currently sought to be answered by AI tools for people eager for answers. They say information is power, and now there is no need to second-guess where that power lies. But that very power requires huge amounts of electrical power to empower AI data centres.
The International Energy Agency (IEA) has estimated global electricity consumption at 415 TWh for data centres in 2024. This figure is likely to double by 2030. Water demand for cooling will also go up exponentially. This soaring demand for data centre-related power and water, coupled with mining of critical minerals to support the hardware for computer infrastructure, is leading to environmental and equity concerns across the globe. Some recent studies have also thrown up worries of a steep temperature increase around data centres.
AI's observed tendency of bias with the use of non-inclusive data is more beneficial to the host countries. It is also increasingly vulnerable to cyberattacks, data leaks, and data misuse, raising huge privacy concerns. Opaque algorithms followed are adversely impacting public trust, amidst fear of information manipulation.
But indeed, there is hope. The voice for ethical validation of AI has been growing stronger. The need for regulatory oversight, including setting lower trajectories for Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE), is now well recognised.
The final solution may arrive somewhere between the EU's strong regulation and the USA's almost nil approach, without causing much headwind to the spirit of innovation.
Falling prices of green energy and energy storage are making that energy route the ready solution for energising data centres. Modular nuclear is evolving into commercial technology, allowing isolated data centre powering. Reverse osmosis and recycling are being perfected to cater to the quality and quantity needs for water. With investments acquiring a new scale, doors are being opened for the advent of technologies like quantum computing, stable thermonuclear reactors, and cooler, high-efficiency data centres requiring less power and water. Necessity, as the mother of invention, is playing out. But the debatable question is whether emerging technologies will make breakthroughs quickly, or whether, by then, less-than-hoped returns will drive away investments in AI, in turn reliving the dotcom nightmare.
India, with a pool of institutions connected to strong data science, physics, mathematics, and statistics, can show the world a more humane, attainable path for AI development. Putting moderate AI regulation in place, building partnerships with countries possessing critical minerals, and creating economic zones for processing with attractive policy could make India an important player in the sector. Public investment needs to be consciously directed towards the sector, incentivising safe, ethical lifecycle AI. Today, Google's largest investment in a data centre outside the USA is all set to come up near Visakhapatnam. These investments are expressions of trust in India's prowess.
If AI fails to pass the tests of ethics, equity, and safety, NPT-like treaties will take off to prevent rogue AI intrusion. But that carries the risk of exacerbating asymmetry among nations. The world needs to come together to facilitate the development and use of responsible, ethical, humane AI.
Views expressed are personal. S Bhattacharya is Chairman of WEBEL and former Chairman of Coal India Limited, and DN Prasad is a Former Adviser, Ministry of Coal
Published by HT Digital Content Services with permission from Millennium Post.