Fresh Delhi – OpenAI CEO Sam Altman addressed growing concerns about the environmental impact of artificial intelligence this week, arguing that while energy consumption is a valid concern, the narrative often overlooks the significant energy demands of human intelligence itself. Speaking at an event hosted by The Indian Express during a major AI summit in India, Altman pushed back against claims of excessive water usage and reframed the discussion around AI’s overall energy efficiency.
Altman was quick to dismiss widely circulated reports about AI’s water footprint, stating that concerns are “totally fake,” though he acknowledged that water was a factor when data centers relied on evaporative cooling. “Now that we don’t do that, you spot these things on the internet where, ‘Don’t use ChatGPT, it’s 17 gallons of water for each query’ or whatever,” he said. “This is completely untrue, totally insane, no connection to reality.” The focus, he argued, should be on the total energy consumption of the rapidly expanding AI landscape.
The debate around AI’s environmental impact has intensified as data centers – the physical infrastructure powering these technologies – grow increasingly energy-intensive. Rising electricity prices have also been linked to the growth of data centers, prompting scrutiny from scientists and policymakers. Currently, there is no legal requirement for tech companies to publicly disclose their energy and water usage, leaving independent researchers to fill the gap.
Altman, in conversation referencing a prior discussion with Bill Gates, disputed claims that a single ChatGPT query consumes the equivalent of 1.5 iPhone battery charges, calling the assertion “There’s no way it’s anything close to that much.” He further contended that comparing the energy required to train an AI model to the energy cost of a single human inference query is “unfair.”
The Human Energy Cost
Altman took the argument a step further, suggesting that the energy expenditure required to develop a human intellect is far greater than that of AI. “But it also takes a lot of energy to train a human,” he said. “It takes like 20 years of life and all of the food you eat during that time before you get smart. And not only that, it took the highly widespread evolution of the 100 billion people that have ever lived and learned not to get eaten by predators and learned how to figure out science and whatever, to produce you.”
He proposed a more equitable comparison: “If you ask ChatGPT a question, how much energy does it take once its model is trained to answer that question versus a human? And probably, AI has already caught up on an energy efficiency basis, measured that way.” This perspective aims to reframe the discussion, highlighting the inherent energy demands of biological intelligence.
Altman’s comments reach as OpenAI continues to expand its presence in India, a key market with roughly 100 million weekly ChatGPT users, second only to the United States. The company opened an office in New Delhi last August and is actively pursuing partnerships to accelerate AI adoption across various sectors.
The Path Forward: Renewable Energy
Despite downplaying the per-query energy cost, Altman acknowledged the importance of addressing the overall energy consumption of AI. He believes a rapid transition to renewable energy sources – nuclear, wind and solar – is crucial to mitigating the environmental impact of the technology. This sentiment aligns with growing calls for sustainable practices within the tech industry.
The competition between OpenAI and Anthropic, led by Dario Amodei, is also playing out in India, with both companies vying for market share and influence. The two CEOs recently made headlines for a notably cool interaction at the AI Impact Summit, choosing not to hold hands during a group photo with Indian Prime Minister Narendra Modi, a moment widely interpreted as a symbol of their intensifying rivalry as reported by the Observer.
As AI continues to permeate more aspects of daily life, the debate surrounding its environmental sustainability will undoubtedly intensify. The focus will likely remain on developing more energy-efficient models and accelerating the transition to cleaner energy sources to power the growing demands of this transformative technology.
What steps will governments and tech companies take to ensure the sustainable development of AI? Share your thoughts in the comments below.
