Sam Altman, the chief executive of ChatGPT creator OpenAI, has defended the vast use of energy by artificial intelligence by comparing it with the food eaten by humans.
AI systems such as ChatGPT use large amounts of energy, both when they are trained and then when they are used to answer queries. The heavy energy use of such systems has been recognised by people including Mr Altman himself, who has suggested that the world needs to adopt new technologies such as solar power and nuclear fusion to power AI in more clean ways.
But in comments made in India during a trip for the country’s AI summit, Mr Altman appeared to dismiss criticisms over that energy use by comparing it to the amount of time and resources used by humans to develop.
“People talk about how much energy it takes to train an AI model – but it also takes a lot of energy to train a human,” Mr Altman said. “It takes about 20 years of life – and all the food you consume during that time – before you become smart.”
In the same comments to The Indian Express, he admitted that it was “fair” to be concerned about energy use but suggested the response should be new forms of energy. “We need to move towards nuclear or wind and solar very quickly,” he said.
During the same conference, he had dismissed concerns about the water used to train AI – where cools data centres – by saying it was “fake”. “It used to be true,” he said, but he suggested that claims people should avoid AI because of the water use are “completely untrue” and “totally insane”.
Mr Altman’s remarks prompted widespread criticism, focused mostly on the fact that he appeared to neglect the value of human life, as well as the fact that he had suggested “training” a human was equivalent to that of an artificial intelligence system.
