Sam Altman Rejects AI Power Use Argument With "Totally Insane" Comment, Internet Reacts

OpenAI CEO Sam Altman has reignited the discussion on artificial intelligence's environmental impact, arguing that humans also consume significant amounts of energy. During an event in India, Altman rejected the AI power use argument with a "totally insane" remark as he weighed in on the energy required to train AI models.

"We used to do evaporative cooling in data centres, but 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 for each query' or whatever," Altman said while speaking at an event hosted by The Indian Express. "This is completely untrue, totally insane, no connection to reality."

Scientists all over the world are trying to understand how data centres, especially amid a rapid growth of AI, are linked to energy usage and rising electricity prices. Altman acknowledged the concerns, urging a shift towards nuclear, wind, and solar energy. "The world is now using so much Al is real, and we need to move towards nuclear or wind and solar very quickly," he added.

Also read | Delhi-Based Uber Driver Completes 50,000 Trips, Reveals Amount Earned In One Month

But his comments on the energy consumed by humans sparked debate, with users criticising his comparison as "evil", "dystopian" and "anti-human".

Altman stated that training a human takes about 20 years of life and food, equivalent to the energy used in AI model training. He also claimed AI has already caught up with humans in energy efficiency, especially when considering a single query.

"One of the things that is always unfair in this comparison is people talk about how much energy it takes to train an Al model relative to how much it costs a human to do one inference query. But it also takes a lot of energy to train a human. It takes like 20 years of life and all of the food you eat during that time before you get smart," Altman argued.

"And not only that, it took like the very widespread evolution of the hundred 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 and then you took whatever you you know you took. So the fair comparison is 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 Al has already caught up on an energy efficiency basis measured that way," he added.

Also read | Employee Quits After Ordering Rs 2 Lakh Worth Shoes Using Self-Created 100% Discount Codes

Social media reaction

"It really is one of the most frightening things I've ever seen a techbro say. Like, he literally doesn't seem to understand that human life has value beyond whatever cost/benefit analysis he applies to implementing lines of code," one user wrote on Reddit.

"They're anti-human humans. The most sickeningly evil thing imaginable," said another user.

"I had someone else argue that healthcare should not be for everyone because not everyone is essential. Like, how do you argue for things that enhance the wellbeing of all when your counterpart doesn't think other hunans matter?" a third user said.