GENERATIVE AI'S DIRTY SECRET

Fortifying search engines with ChatGPT-grade smarts could require a fivefold increase in computing power, with the carbon emissions to match.

- CHRIS STOKEL-WALKER

IN EARLY FEBRUARY, Google and Microsoft announced major overhauls to their search engines. Both tech giants have spent big on generative AI tools, which use large language models to understand and respond to complex questions. Now they are racing to integrate them into search, hoping to give users a richer, more accurate experience. The Chinese search company Baidu rushed out its own version of this tech in March.

But the excitement over these efforts conceals a dirty secret. AI-fortified search engines are likely to demand far more computing power, which means a massive increase in energy use and carbon emissions.

"There are already huge resources involved in indexing and searching internet content, but the incorporation of AI requires a different kind of firepower," says Alan Woodward, professor of cybersecurity at the University of Surrey in the UK. "Every time we see a step change in online processing, we see significant increases in the power and cooling resources required by large processing centers. I think this could be such a step."

For starters, training the large language models that underpin OpenAI's Chat GPT which powers Microsoft's souped-up Bing search engine and Google's equivalent, Bard-"takes a huge amount of computational power," says Carlos Gómez-Rodríguez, a computer scientist at the University of Coruña in Spain. "Right now, only the Big Tech companies can train them."