Google Fails to Meet Important Climate Goals, So Its AI May Tell You to Put Glue on Pizza

Maybe you should watch your AI consumption.

Energizer Bunny

Google’s latest environmental impact report shows the Silicon Valley stalwart is woefully behind schedule on its ambitious plan to achieve net-zero carbon emissions by 2030, as reported by The Associated Press. The culprit? AI.

According to APGoogle’s carbon emissions in 2023 saw a staggering 13% annual increase from the previous year and 49% from 2019, with Google citing its AI efforts as the driving force behind this change.

While AI’s energy requirements remain largely invisible to the public eye, it’s an incredibly resource-intensive field. AI models consume virtually all of their energy (a search for OpenAI’s ChatGPT, for example, is estimated to be the equivalent of ten Google searches) by straining the data centers that run the models. The servers in those data centers are also at risk of overheating, meaning companies like Google and Microsoft often have to use water to cool them. (Google and Microsoft both use air cooling as long as outside temperatures stay below a certain threshold, but in a warming world, there are and will continue to be plenty of hot days.)

In short, Google’s quest to stay in Silicon Valley’s AI race has come at the expense of its efforts to reduce carbon emissions. And while Google and other AI makers continue to promote AI and machine learning as tools to mitigate climate change, the public has mostly seen generative AI products like Google’s AI-embedded search telling us to eat rocks and put glue on pizza.

What can I say? At least the cheese won’t slide off our slices while the planet is dying.

Don’t be mean

As stated in its report, Google has has made an effort to integrate renewable energy sources into its overall energy infrastructure. Despite this, the fact remains that the company is moving further and further away from its climate target, even though am Kate Brandt, Google’s chief sustainability officer, told the APIts net zero emissions mission is a very difficult task anyway.

“Achieving this goal of net zero emissions by 2030,” Brandt told the AP“is an extremely ambitious goal.”

“We know this will not be easy and that our approach will need to continue to evolve,” she continued, “and this will require us to manage many uncertainties, including the future of AI’s environmental impacts.”

Meanwhile, experts like Amanda Smith, a senior scientist at the nonprofit Project Drawdown, are urging everyone — AI creators and everyday users alike — to be mindful of their AI consumption.

“It’s up to us humans to monitor what we do with” AI, Smith told the APadding that “when it is worthwhile, we can ensure that these demands will be met by clean energy sources.”

Learn more about Google and AI: Google uses an incredible amount of water for its AI