-
AI already has a massive carbon footprint
-
This is only going to grow as AI gets embedded in everything
-
Continued training of enterprise AI models will probably be a massive energy hog over time
Artificial intelligence (AI) is fast becoming as ubiquitous as the cloud for both enterprise and consumer users. However, the AI boom is going to lead to major environmental concerns about AI as its usage grows because this is yet another technology that burns through electricity apace, Hussein Hallak, co-founder of AI art firm Momentable told Silverlinings.
Data centers for the cloud, blockchain and other technologies have already raised the energy level needed for technology at a time when the environment can least afford it. Now AI will increase this massive carbon footprint even further as the number of people that use AI dramatically increases, Hallak told us.
This is everything from the tiniest device to the most massive data center. “At a certain point AI we will be embedded in everything,” including data centers, he asserted.
So what are the biggest power hogs in AI right now? The AI models themselves — particularly unrefined or untrained models — use an awful lot of power since they require massive server farms to implement deep learning training, he said.
As Nvidia notes in a blog: “A popular way to get started with deep learning is to run these frameworks in the cloud. However, as enterprises start to grow and mature their AI expertise, they look for ways to run these frameworks in their own data centers, to avoid the costs and other challenges of cloud-based AI.”
Which all sounds fine and dandy, but as Hallak noted, when companies start to train models on their data, “they will use an enormous amount of data that they have to train on, and continue to train on,” he said.
He suggested that enterprise AI use could be biggest energy consumption factor in the future.
Who's got the power?
Where the power to drive AI is coming from is maybe one of the biggest factors in this whole conundrum. Is it from fossil fuels or renewable energy sources like wind and solar? It depends on the region of the world, according to Hallak.
“The least worry is from areas like Europe and the U.S.,” he claimed. “The biggest worry is counties like China and Russia, because in their use of energy, they don’t have the same guardrails and they’re trying to catch up...they don’t really care if they’re using fossil fuels.”
Of course, we could argue about rampant U.S. and European use of various types of fossil fuels, but that’s not really within the scope of this simple story. There’s certainly enough blame to go around!
“I think that unless we solve the main problem at its source, which is how we generate energy, pointing at certain technologies is not going to help at all,” Hallak states.
So, no problem, right?
Want to discuss AI workloads, automation and data center physical infrastructure challenges with us? Meet us in Sonoma, Calif., from Dec. 6-7 for our Cloud Executive Summit. You won't be sorry.