Why is it making climate-conscious users anxious?

However, for Gen Z and millennials, who prefer paper straws and paper bags to curb their carbon footprint, the shift in the digital world feels like a moral dilemma.

Young people are encouraged to adopt greener lifestyles while being introduced to digital systems that produce a mostly invisible environmental footprint.

The climate crisis has been one of the top concerns for millennials and Gen Zs globally. Deloitte Global’s 2025 Gen Z and Millennial Survey found that roughly two-thirds of both cohorts felt anxious about climate change and were willing to pay more to purchase environmentally sustainable products and services.

“I should be able to choose. I didn’t ask for AI-generated search results,” says Delhi-based Agnes Thomas, 28, a digital marketing manager. “By choice, I would not want to work with AI at all.” Yashita Gupta, a lawyer and climate educator who co-founded Prakriti Bodh, an initiative introducing climate education to children, concurs. “Even the simplest websites come loaded with AI be it Zoom or email. I never agreed to this.”

Power guzzler

AI automation promises speed and ease, but at a growing ecological cost. Since 2019, Google’s carbon emissions have increased by 51%, driven largely by the energy demands of its AI infrastructure.

Training a single large model like GPT-3 can emit as much carbon dioxide as five American cars do over their entire lifetimes, according to an MIT Tech Review report.

Google’s own sustainability data shows that its data centres consumed more than 5.6 billion gallons of water in 2022 used mostly to cool AI servers. This is equivalent to the annual water intake of 15-20 million people. Apart from water, electricity consumption by data centres is also going to take up a substantial share—4.5% of the global consumption by 2030, according to SemiAnalysis.

“AI is becoming one of the most resource-intensive parts of the tech sector,” says Shaolei Ren, a researcher on AI and energy at the University of California, Riverside. “Its emissions are on track to rival, even exceed, the public health impacts of road transport emissions in major US states by 2030.”

AI and India

In countries like India, the stakes are even higher. The AI infrastructure boom is unfolding atop already fragile energy and water systems. The country’s data centre capacity is expected to grow 77% by 2027 from the current levels, reaching 1.8 GW, according to global real estate services firm JLL.

Metropolitan cities like Mumbai, Bengaluru, Delhi, Chennai have emerged as key hubs, many of which are already grappling with water scarcity. “We’ve seen this elsewhere,” says Ren. “In Chile, Spain, and Arizona, communities have protested against data centres hoarding water during droughts.”

However, despite the risks, the usage of AI has become a part of life globally. Nearly 80% of ChatGPT users in India say they use it on a daily or weekly basis. As of mid-2025, ChatGPT alone processed over 2.5 billion prompts per day, a 150% jump from late 2024.

Around 35% of these prompts are for work or productivity tasks, 22% for coding, 18% for education, and 15% for content creation. According to Dr. Mohammed Atiqul Islam, assistant professor of computer science and engineering at the University of Texas, with AI addressing millions of queries from millions of users, inference (process input data and generate an output) ends up dominating the total energy consumption at the data centres.

Opting out is increasingly difficult as AI usage becomes ubiquitous and institutionally embedded. In 2024, 78% of organizations globally used AI in at least one business function, up from 55% the year before, according to a report by McKinsey and Company in March this year.

In India alone, 87% of GDP sectors are already using AI at the ‘enthusiast’ or ‘expert’ level, a 2024 Nasscom report showed. In fact, AI adoption rate is the highest in the country at 59%, closely followed by the UAE (58%), an analysis by Forbes Advisor India showed.

AI is also becoming increasingly entrenched in the public sphere, driven by the Indian government’s growing adoption of the technology. The AI Competency Framework launched in April 2025 is now training over 3.1 million civil servants in ethical and context-aware AI deployment.

AI-enabled MyGov Helpdesk on WhatsApp handled over 36 million citizen queries during Covid, while state-level AI chatbots like IRCTC’s AskDISHA and NPCI’s PAi are now embedded in public service delivery. Yet, only five Indian states have formal AI policies, exposing governance gaps despite widespread usage.

What’s the fix?

Some tech companies are pitching a cleaner path forward with Green AI: the idea of training smaller, more efficient models or powering data centres with renewables. “You can run a data center on solar or wind power with zero carbon and zero water impact. Dry cooling is possible. These setups are more expensive but doable,” says Dr. Islam.

Ren is cautiously optimistic. “Green AI is promising but without regulation or reporting mandates, voluntary efforts won’t be enough. We need the digital equivalent of nutrition labels, that is, standardised disclosures on water and energy use.”

As AI weaves deeper into the fabric of work, services, and governance, the question of choice becomes increasingly urgent. “Whether people have a real choice in adopting AI depends on how the regulatory space evolves,” says Dr. Aakansha Natani, assistant professor at IIT Hyderabad.

“There are two contrasting regulatory approaches. Ex-ante, where you anticipate risks and regulate in advance, like the EU’s AI Act; and ex-post, where you wait for issues to emerge and then respond, like the US model.”

As the AI age dawns, the question remains: Can a generation in a do-or-die climate fight afford to fuel the AI that is making it worse?