To date, the rapid growth of artificial intelligence has proven almost uncontrollable. As the technology has swept through the tech industry like wildfire, regulators have been largely powerless to stay ahead of its proliferation and evolution. Questions about the scope and liability of artificial intelligence abound, but few are answered. Add to that the issue of the industry’s huge and growing energy footprint and associated carbon emissions, which are now so large that developed countries are facing the most severe energy crisis they have seen since the shale revolution.
“AI-powered services require significantly more computing power, and therefore electricity, than regular online activity, sparking a series of warnings about the technology’s environmental impact,” the BBC recently reported. A recent study by Cornell University scientists found that generative AI systems like ChatGPT use up to 33 times more energy than computers running task-specific software, and each AI-powered internet search consumes around 10 times more energy than a regular search.
The global AI sector is expected to account for 3.5% of global electricity consumption by 2030. In the United States, data centers alone could consume 9% of electricity production by 2030, double what they are today. This development is already sending ripples through Big Tech: earlier this month, Google revealed that its carbon emissions have soared 48% over the past five years.
Not only does the United States need a massive increase in renewable energy to keep up with the insatiable demands of the technology sector, it also needs to ramp up energy production even more to avoid serious shortages. Slowing runaway AI energy consumption will require widespread and rapid action on multiple fronts, but the United States’ own national security concerns require it to keep up with other nations’ AI spending and development. The genie is out of the bottle, and there’s no putting him back in.
“Certain strategic areas of U.S. government artificial intelligence capabilities currently lag behind industry while foreign adversaries are investing heavily in AI,” a recent Department of Energy (DoE) bulletin stated. “Without rapid U.S. government leadership in this area, the nation risks falling behind in the development of safe and trustworthy AI for national security, energy, and scientific discovery, thereby undermining our ability to address pressing domestic and global challenges.”
Therefore, the question now is not how to prevent a global AI takeover, but how to fast track new energy sources, how to place strategic limits on the rate of growth and consumption in the energy sector, and how to ensure that AI is used responsibly for the benefit of the energy sector, countries, their people, and the world at large.
To that end, the U.S. Department of Energy (DoE) has proposed a new agency-wide effort to “harness and advance artificial intelligence for the public good,” Axios reports. This month, DoE released a roadmap for the program, which was first mentioned publicly in May of this year. Frontiers in Artificial Intelligence for Science, Security, and Technology (FASST) involves coordinated efforts by all 17 of DoE’s national laboratories.
The program will be focused on remaining competitive in AI on a global scale, but will also devote significant resources to creating more energy-efficient computer models to ensure the nation’s energy security and climate goals are not compromised in the process. The program’s five overarching goals are:
1. Advancing national security 2. Recruiting and developing talented people 3. Leveraging AI for scientific discovery 4. Solving energy challenges 5. Developing the technical expertise needed for AI governance
Under the goal of “Addressing our Energy Challenges,” the Department of Energy states, “FASST will unlock new clean energy sources, optimize energy production, improve grid resiliency, and build the advanced energy economy of tomorrow. America needs low-cost energy to support economic growth, and FASST will help meet this challenge.”
The proposed FASST program would be an important step in the right direction toward the responsible growth and application of artificial intelligence in the United States, but it still needs Congressional approval and funding before it can take effect. A bipartisan bill has already been introduced in the Senate.
Article by Haley Zaremba of Oilprice.com
Other popular articles from Oilprice.com:
