AI's Massive Power Consumption Demands Innovative Energy Solutions

We're only just discovering the depths of AI's potential, but it already seems impossible to imagine a future without it. Love it or hate it, stunning technological breakthrough or new world order, it seems to be a given that artificial intelligence is here to stay. But it turns out that the continued advancement of AI is not a guarantee unless an energy breakthrough occurs that would allow us to meet its insatiable demand for electricity. Reuters reports that artificial intelligence "will consume vastly more power than people have expected" and meeting that need will require as-yet unproven and undiscovered technological advances.  

Already, the carbon footprint of Artificial Intelligence is almost as large as that of Bitcoin - which is to say, it consumes as much energy as some entire developed nations. "Currently, the entire IT industry is responsible for around 2 percent of global CO2 emissions," Science Alert reported last year. And it's set to grow at a breakneck pace. Consulting firm Gartner projects that if business continues as usual, the AI sector alone will consume 3.5 percent of global electricity by 2030. 

Estimations suggest that the training process for GPT-3, the predecessor of ChatGPT, required around 1,287 megawatt hours of electricity and 10,000 computer chips - enough energy to power about 121 homes in the United States (where the average home consumes a considerable amount of energy compared to other nations) for an entire year. The amount of energy used to train GPT-3 is also sufficient to produce around 550 tonnes of carbon dioxide. Indeed, estimated figures suggest that Open.AI, the creators of ChatGPT, spend about US$700,000 per day on computing costs alone to run its chatbot service for its approximately 100 million users around the globe. 

Just this week, OpenAI chief executive officer Sam Altman spoke to the fact that artificial intelligence is rushing headfirst toward a brick wall - while the technology continues to spread and advance at a rapid clip, there's no plan for how its operations will continue to secure enough energy going forward. "There's no way to get there without a breakthrough," he said on Tuesday at a Bloomberg event on the sidelines of the World Economic Forum's annual meeting in Davos. But he said there is a light at the end of the tunel, and it lies in the research and development of carbon-free energy production. In particular, Altman says that the energy puzzle facing the future of AI " motivates us to go invest more in fusion."

Altman is putting his money where his mouth is. In 2021, he personally provided $375 million in funding to private U.S. nuclear fusion company Helion Energy, which has since inked a deal with Microsoft to provide energy to the world's largest software maker. Altman is also chairman of the board for nuclear microreactor company Oklo, which he hopes to help take public this year. The AI systems of the future will need tremendous amounts of energy and this fission and fusion can help deliver them," Altman told the Wall Street Journal last year. 

While AI is poised to consume a problematic amount of energy, it could also be integral to the global decarbonization process. Meeting global climate goals will require systems transformation at an unprecedented scale, which will be impossible without intelligent, responsive, and flexible computing systems able to recognize and predict complex patterns of production and consumption. Already, AI is integral to renewable energy forecasting, smart grids, coordination of energy demand and distribution, maximizing efficiency of power production, and research and development of new materials. 

AI may be the key to beating global warming - if it's used and powered responsibly. "Fundamentally speaking, if you do want to save the planet with AI, you have to consider also the environmental footprint," Sasha Luccioni, researcher of ethics at the open-source machine learning platform Hugging Face, was quoted by The Guardian. "It doesn't make sense to burn a forest and then use AI to track deforestation."

By Haley Zaremba for Oilprice.com

More Top Reads From Oilprice.com:

• Copper and Gold: Key Indicators for Predicting Economic Trends
• The Strategic Implications of Iran's Recent Missile and Drone Attacks
• Canada's Uranium Is Fueling the World's Nuclear Energy Boom