A new report has identified a spike in the electricity consumption rates of artificial intelligence (AI) chip manufacturing by 350% since 2023.
The research paper, published by Greenpeace East Asia, notes that East Asia bears the brunt of the climate impact of AI chip manufacturing. Asian countries like Taiwan and South Korea are leading hubs for semiconductor manufacturing but soaring electricity demand is denting climate change targets.
Semiconductor manufacturing plants in the region are turning to fossil fuels to power their operation. The combined global emissions from AI chip manufacturing in Asia Pacific, particularly South East Asia, could cause the region to miss out on its environmental sustainability targets.
“The manufacturing process of AI hardware is energy intensive and carries a significant environmental footprint, especially considering the concentration of this manufacturing in East Asia, where power grids still rely heavily on fossil fuels, and chipmakers have taken few steps to procure renewable energy,” read the report.
Between 2023 and 2024, electricity consumption for AI chip manufacturing soared by over 350%, threatening national electricity grids in the region. Given the demand and pace of adoption, the report tips electricity consumption rates to rise by 170-fold by the end of the decade.
At this stage, the electricity consumption rates will match the total consumption of Ireland, outstripping the consumption rates of smaller countries. Currently, countries in the Asia Pacific are unfurling new fossil fuel plants to power their semiconductor activities with Taiwan leading the charge.
The trio of South Korea, Japan, and Taiwan have since launched LNG plants, justifying the boom of AI and semiconductor industries as a reason for the expansion.
Turning to renewable energy remains possible
While fossil fuels are an easy alternative for chip manufacturing countries, the report highlights the danger it poses to the immediate environment and long-term climate change targets.
Aware of the raging demand for AI chips, Greenpeace urges an industry-wide commitment to embrace renewable energy sources. It said hardware manufacturers can invest in wind and solar power.
The report namechecks U.S.-based chip companies like Nvidia (NASDAQ: NVDA) and AMD (NASDAQ: AMD), noting that the mad dash for revenue adversely affects the environment. Research analysts say that Nvidia and AMD should take responsibility for the impact of their supply chains in Asia.
“It is crucial for Nvidia, AMD, and their peers to fully recognize the environmental impact of their supply chains and work with manufacturers to increase renewable energy use,” read the paper.
Experts criticize AI-written paper over climate change denial
Researchers are poking holes into an AI-generated paper denying climate change under the illusion of objectivity.
According to a report by Fact Check, the AI-generated paper rejects popular and peer-reviewed models on human-induced global warming. The paper titled “A Critical Reassessment of the Anthropogenic CO2-Global Warming Hypothesis” gained traction among global warming deniers.
The paper claims to be generated solely by Grok 3 AI, garnering over a million views on X (formerly Twitter). Researchers are taking swipes over the paper’s claim of neutrality, noting that the illusion of objectivity is non-existent in large language models.
“Large language models do not have the capacity to reason,” said environmental science professor Mark Neff. “They are statistical models predicting future words or phrases based on what they have been trained on.”
Climate researchers say the paper received input from a raft of co-authors “guiding its development.” A close look at the co-authors reveals a pattern of climate change denial, with contrarian author Willie Soon leading the cohort after receiving over 1 million dollars from a key player in the fossil fuel industry.
Other co-authors have seen their papers scientifically contested for their claims against human-induced global warming. Apart from leaning on research from climate change contrarians, experts criticize the AI-generated paper for failing to include the prompts used to generate the results.
“It includes datasets that formed the basis of the paper, but no prompts,” Microbiologist Elisabeth Bik said. “We know nothing about how the authors asked the AI to analyze the data.”
Furthermore, others are saying that the claim that it is entirely AI-generated is worrisome given the unverifiable nature of its objectivity. A statement from Grok’s official X account denied its involvement in creating the paper, criticizing it over its simplified assumptions and agenda-driven approach to research.
A murky review process is the final blow, with experts pointing out that the AI did not pass through the crucible of either an open or double-blind review.
AI risks reach a tipping point
Experts have raised alarm over the dangers of emerging technologies and AI. A World Economic Forum (WEF) report cites the potential for misinformation by the indiscriminate use of generative AI as a major area of concern for regulators.
Deepfakes are running riot in several jurisdictions, forcing authorities and technology companies to create stricter regulations for consumer safety. The potential for LLMs to make errors is putting healthcare and pharmaceutical industries on high alert as the World Health Organization warns for a cautious approach.
In order for artificial intelligence (AI) to work right within the law and thrive in the face of growing challenges, it needs to integrate an enterprise blockchain system that ensures data input quality and ownership—allowing it to keep data safe while also guaranteeing the immutability of data. Check out CoinGeek’s coverage on this emerging tech to learn more why Enterprise blockchain will be the backbone of AI.
Watch: Artificial intelligence needs blockchain
title=”YouTube video player” frameborder=”0″ allow=”accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share” referrerpolicy=”strict-origin-when-cross-origin” allowfullscreen=””>