Sam Altman would like to remind you that humans use a lot of energy, too

Sam Altman, CEO of OpenAI, addressed concerns about the environmental impact of AI this week while speaking at an event hosted by The Indian Express.

For one thing, Altman — who was in India for a major AI summit — said concerns about AI using water were “completely false,” though he acknowledged it was a real problem when “we were used to evaporative cooling in data centers.”

“Now that we don’t do that, you see these things on the Internet where, ‘Don’t use ChatGPT, it’s 17 gallons of water per query’ or whatever,” Altman said. “This is completely untrue, completely crazy, and has nothing to do with reality.”

He added that it is “fair” to be concerned about “energy consumption – not per query, but in aggregate, because the world is now using a lot of artificial intelligence.” In his view, this means that the world needs to “move towards nuclear or wind and solar very quickly.”

There is no legal requirement for technology companies to disclose how much energy and water they use, so scientists are trying to study it independently. Data centers have also been linked to rising electricity prices.

Citing a previous conversation with Bill Gates, the interviewer asked if it was accurate to say that a single ChatGPT query currently uses the equivalent of 1.5 iPhone battery charges, to which Altman replied: “There’s no way it’s anywhere near that much.”

Altman also complained that many discussions of ChatGPT’s energy use are “unfair,” especially when they focus on “the amount of energy needed to train an AI model, compared to the cost of a human running a single inferential query.”

“But training a human also requires a lot of energy,” Altman said. “It takes about 20 years of life and all the food you eat during that time before you become intelligent. And not only that, it took very large-scale evolution of the 100 billion people who have ever lived and learned not to be eaten by predators and learned how to figure out science and everything else, to produce you.”

So, from his perspective, a fair comparison is: “If you ask ChatGPT a question, how much energy does it consume once it trains its model to answer that question versus a human? And the AI ​​has probably already achieved the baseline of energy efficiency, and it’s measured that way.”

You can watch the full interview below. The conversation about water and energy use begins at around 26:35.