Eighteen months ago, the federal government was just beginning to grapple with the rise of artificial intelligence. The landscape was evolving quickly in the months after OpenAI’s release of ChatGPT in November 2022, raising questions about how the U.S. would stay ahead of the competition — and minimize risks.
In October 2023, President Joe Biden signed Executive Order 14110. In its 111 pages, the order outlined how to “ensure that America leads the way in seizing the promise and managing the risks of artificial intelligence.”
It covered every corner of the administration’s governance of AI. And of course, that includes the energy sector.
As a result of the order, the Department of Energy got a new office. Dubbed the Office of Critical and Emerging Technology, it pulls together the efforts of 17 national laboratories and other DOE research programs on critical technologies — biotechnology, quantum, microelectronics, and semiconductors in addition to AI — to inform policymaking, and set up partnerships with both industry and academia.
In the last year, the office’s research has included a report on opportunities for AI to help the grid, such as in planning and operations, as well as a proposed initiative to “build the world’s most powerful integrated scientific AI systems,” called Frontiers in Artificial Intelligence for Science, Security and Technology (or FASST).
But since the election of Donald Trump, the office’s future is in question. On the one hand, Trump has said he would revoke Biden’s executive order, leaving CET and other AI-focused federal efforts in limbo. But on the other, he has made leadership in AI a key part of his platform — which is arguably what CET is aiming to do as well.
At Latitude Media’s Transition-AI event this week, I spoke with the office’s director Helena Fu (who also serves as DOE’s Chief Artificial Intelligence Officer) and policy advisor Charles Yang, to take stock of DOE’s energy efforts. Below, find excerpts of our conversation, edited for brevity and clarity.
Can you characterize what the mood is at DOE and specifically within CET during this lame duck period? How are you thinking about the possibilities here?
Helena Fu: It was actually in the first Trump administration that DOE created an AI technologies office. So this office is more the growth of that prior effort to technologies beyond AI. It’s a recognition that there are other foundational technologies that DOE is deeply engaged in that also need to be brought to the fore. These technologies are incredibly important for our economic prosperity, for our national security.
Right now, [the mood at the office includes both] urgency and purpose, because DOE sits at this intersection of AI and its potential applications in the energy sector — and also energy for AI. And both of these things are going to be incredibly critical in both the short-term and the medium- and long-term years ahead.
Charles Yang: I don’t know about the rest of DOE, but I’ll just say we’re all reading the same news [suggesting that Trump wants to roll back Biden’s executive order].
But I think one thing that’s exciting is seeing the incoming administration recognize the critical role that energy plays in our national security and in our economic development, with the creation of this new National Energy Council. We’ve seen that both data centers and the broader onshoring of manufacturing have significant loads. And our ability to meet that load is part of both our national security mission, but also just part of our broader economic engine. And so all this is to say that there’s a lot of renewed interest in energy and in the work that the department does, which I think can only be a good thing for us.
Is there something that you think the industry at large generally thinks about incorrectly, either about AI as a whole or specifically about DOE’s work on AI?
Charles Yang: So I wouldn’t necessarily say “incorrectly,” but I think the value and the reason DOE needs to be doing this work is because industry is focused on very specific kinds of AI models, specifically the large language models.The advances they’ve made are certainly tremendous and really exciting. But they’re not necessarily the end all, be all.
One thing I’m really excited about is we have these new AI models that are able to make predictions and inferences at a scale that is far more rapid than we’ve seen before. But these autonomous experimentation platforms require robotics. They require coordination of infrastructure, scientific equipment, and AI. And stitching all of that together is work that is going to be done by the public sector. These models that are important, particularly for scientific domains, are exactly the kinds of ones that I think we see industry has not quite invested as heavily in for the domains and applications that we care about.
And so this is not to say that we are competing with industry, but simply we’re focused on prioritizing different kinds of models and different kinds of applications. But we’re certainly working closely with them in finding areas of common overlap.
Do you have any reactions to what we’ve heard from the Trump transition team so far — specifically related to this Manhattan Project for AI concept, but also more generally to the National Energy Council — and how are you thinking about the transition that the Department of Energy needs to make?
Charles Yang: I would say one observation is just that if Chris Wright is confirmed, he would be the first engineer to be the Secretary of Energy in arguably 30 years. And I think that’s an important recognition of the role that technology plays in our national security.
I think when you see these calls for a Manhattan Project for AI, it [evokes an understanding that] we need this public sector investment because it’s an important technology. DOE also played a critical role in the original Manhattan Project — it’s really kind of our birth and origin story.
Helena Fu: There are two things that really resonate about the Manhattan Project framing. The first really is scale: the need to move quickly and at scale to do a thing. That is something that DOE as a mission oriented R&D agency is very much focused on. And the second is urgency: urgency to harness AI and unleash its transformative potential to solve problems, and also the urgency to make sure that we are meeting energy demand for AI, but also for manufacturing in the United States. And I think both of those things feel very much along the lines of what DOE is already trying to do and thinking through.
