In its annual environment report released earlier this month, Google reported a 13% increase in its emissions footprint in 2023 compared with the previous year. The rise was attributed mainly to the increased electricity consumption in its data centres and supply chains. Google said its data centres consumed 17% more electricity in 2023, and added that this trend was expected to continue in the coming years because of greater deployment and usage of its artificial intelligence (AI) tools.
Power guzzling intelligence
AI, which is expected to enable transformative changes across several domains, including attempts to find solutions to climate change, has a very heavy emissions footprint, the scale of which is becoming evident only now.
Studies have shown that a simple AI query, like the ones posted to Openai’s chatbot CHATGPT, could be using between 10 and 33 times more energy than a regular Google search. Image- based AI searches could be using even more energy.
Why emissions are higher
AI models typically work much more than a simple Google search even when the same question is addressed to both. They sift through much more data while processing and formulating appropriate responses. More work means a greater number of electrical signals are required when the computer is processing, storing, or retrieving data.
More work also generates and releases more heat, which then requires more powerful air- conditioning or other forms of cooling in the data centres.
A worrying prognosis
As AI tools are deployed more widely, their impact on energy consumption worldwide is expected to rise sharply. Already, data centres account for between 1% and 1.3% of the global electricity demand. This could double (become between 1.5% and 3%) by 2026, according to recent projections of the International Energy Agency (IEA). By contrast, despite the large number of electric vehicles on the road, their share of global electricity consumption was just about 0.5%, the IEA said.
At the level of countries, the electricity consumption of data centres as a share of the national demand has already crossed double digits in several regions.
In Ireland, which has a disproportionately large number of data centres because of the tax breaks and incentives it offers, this share has reached 18%, IEA numbers show. In the United States, the country with the largest number of data centres, this number was estimated to be between 1.3% and 4.5%. The numbers for India were not available.
Scenario for India
Saurabh Rai, CEO of Arahas Technologies, which works in the sustainability space, said the situation in India, at least in the next few years, was not likely to change much, and the huge environmental toll of AI and data centres would become evident very soon.
“It is not just about electricity consumption. There is an increased demand on water resources as well, required for cooling of data centres. There is inadequate data on water consumption of data centres but the centre that serves Openai’s GPT- 4 model in Iowa (US) is reported to have consumed 6% of the district’s water supply in July 2022,” Rai said.
Rai said these had implications for India, where the deployment of AI and data centres is expected to increase rapidly in the coming years.
“AI is revolutionary technology, and we will see it being used widely in India. But the emerging trends about its environmental impact means that we should plan its expansion in a manner that minimises the adverse impacts. Simultaneously, we need to make sure that the companies running these data centres take every measure to make their processes efficient and minimise the emissions footprint,” he said.
Alternative view
Other estimates suggest that the large scale deployment of AI could help in significant reductions of emissions globally. A recent study by the Boston Consulting Group found that application of AI to corporate and industrial practices could result in a 5-10% reduction in global emissions by 2030, while generating a value worth $ 1.3 trillion to $ 2.6 trillion through additional revenues or cost savings.
Emissions reductions can happen if AI is deployed to monitor and predict emissions in existing processes, and optimise these to eliminate wastage or inefficiencies. (Source: The Indian Express)
