What do frozen peas and artificial intelligence (AI) have to do with one another? No, it’s not a new Sci-fi movie, but a way for grocery stores to conserve energy during peak demand times with a bit of help from technology.  The idea being AI can predict when the need will be high and lower freezer temps earlier in the day, thereby autonomously returning to normal temps during peak times.

In the same article, the author notes, “AI is poised to transform the entire energy sector in the coming years, by helping overcome energy’s inherently variable and uncertain nature, and by accelerating the adoption of renewables.”

It’s not just grocery stores making the most of AI to conserve energy. Google uses the technology to predict when its data centers get too hot, so that cooling systems are automatically activated, reducing energy costs by 40 percent. Manufacturing companies, oil and gas firms, energy services, transportation, forestry, the environment, and many other industries also use smart technology to drive efficiencies and reduce energy consumption.

Smarter energy storage makes it easier to harness power.

There’s tremendous pressure on utilities to have storm preparedness plans, so they use outage management systems for grid oversight and restoring power during outages. Advancements in AI can only improve these platforms.

According to BCG Analysis, AI for climate control could help reduce five to ten percent of greenhouse gas emissions by 2030. Remember the 2021 Texas winter snowstorm that caused excessive power outages? AI could have predicted the weather, prepared for the outages by activating alternative energy sources, and routed power to locations that needed it most.

It takes a lot of power to run machines.

However, concerns surround AI, including how much computing power is needed to run the complex algorithms, not to mention the “training” and “building” of AI models.  Some studies compare building them to automobile emissions and powering homes. So, while AI has the “power” to accelerate reductions in global carbon emissions, nearly 300,000 kilograms of carbon dioxide equivalent emissions are created during training a single model, which is the equivalent of 125 round-trip flights between New York and Beijing. That’s because a lot goes into AI, including cloud storage, hardware, and server locations. It also takes a lot of attempts to perfect the algorithms, which in turn means energy consumption.

According to the EPA, in 2019, electricity consumption, including fossil fuels and natural gas (25 percent), was second to transportation (29 percent). Work is being done, however, to lower AI’s carbon footprint. MIT researchers, for example, have implemented a new automated AI system for training and running specific neural networks.  According to the article “Reducing the Footprint of Artificial Intelligence,” researchers could reduce AI’s carbon byproducts to the low triple digits.

Is renewable the answer?

Aside from relying on a hundred percent renewable energy, other solutions to lowering AI’s carbon footprint may include reducing the “training” process. The movement, Green AI, aims to make machine learning cleaner, moving training sessions to locations that use renewable or green energy. There’s even a tool to measure the electricity an AI program will use and its equivalent carbon output.

In the end, AI plays a pertinent role in improving power grids and a host of other industries, and there is ongoing research to improve the energy that goes into it.

At FootBridge, our recruiters specialize in finding the energy contract talent our clients need, so you’ll always have experienced professionals ready to step into any project. Contact us today to learn more.