Nuclear renaissance: JFAC visits INL, learns of artificial intelligence’s place in nuclear power


IDAHO FALLS — Idaho’s Joint Financial Appropriations and Finance committee made their last stop at Idaho National Laboratory Wednesday after a three-day tour of Idaho Falls.

The tour took Idaho senators and representatives on the committee to the College of Eastern Idaho, local organizations like the Development Workshop and Safety, Prevention and Resource Center, and lastly, INL.

RELATED | CEI hosts JFAC in historic visit to the community college

RELATED | EIRMC CEO, other medical leaders say Idaho Legislature not doing enough to support medical workforce

At INL, one of the talks that sparked interest among members of JFAC was a presentation on the laboratory’s concerns and uses of artificial intelligence.

Chris Ritter, director of INL’s Digital Innovation Center of Excellence for scientific computing and AI, discussed how fast the technology has changed since the introduction of AI a few years ago.

Ritter said many years ago that AI services like ChatGPT, given prompts, never produced accurate results.

During the presentation, officials showed an image generated by an AI prompt several years ago of a woolly mammoth. While it got most of it right, the AI generation had five legs and was missing a tusk. However, given the same prompt today, the same AI services can create GIFs of accurate-looking mammoths walking around.

“You can now add imagery and make things up that didn’t happen,” Ritter said. “From a national security perspective… we’re seeing those kinds of videos getting out there.”

Ritter and those within INL have concerns about the possible uses of AI, which can be used for good or for harmful reasons.

“It’s incredibly important that America is the leader when it comes to AI,” Ritter said.

Currently, the United States is behind in AI technology compared to China, which he said is on track in ensuring its AI facilities are fully powered. Ritter said, “We need power.”

“Nuclear is a key ingredient to the country, leading in AI and all the things we want to do as a country,” he said.

Members of JFAC at INL on May 21. | Courtesy INL
Members of JFAC at INL on May 21. | Courtesy Idaho National Laboratory

Officials said the concern is that China is building infrastructure to support 32 gigawatts of power, while similar efforts aren’t being made in the US. Taking testimony from Eric Schmidt, former chairman of Google, Ritter said the country will need 29 gigawatts by 2027 and 67 more gigawatts by 2030.

“It’s roughly 100 gigawatts of more energy that we’re not producing,” Ritter said.

While the country hasn’t created more power, he said we’ve become more efficient. However, the relationship between power and processing ability with AI requires more energy.

Discussing China’s Deepseek AI, Ritter said that aside from the hit in the stock market, there is concern over the quick adoption of this AI by other countries.

Ritter said the US is attempting to pull ahead by funding more nuclear reactors to meet the power requirement for AI. Meta, Amazon, Google and Microsoft are funding major infrastructure projects.

However, one issue affecting all of them is finding and ensuring sufficient fuel and materials for these reactors.

This is where the INL looks to use AI to test whether these materials are suitable for use in reactors after irradiation tests.

Ritter said these very repeatable processes could be taken over by AI and performed 24/7, whereas a worker could only test for 10 hours a day.

“We can get more fuel qualified more quickly, and get more use out of these federal funds we’ve already invested in these facilities,” Ritter said.

The other part of incorporating AI is managing microreactors with it. He said a typical microreactor needs 16 members to operate, but a few would be needed with AI.

“You have this kind of pairing of humans doing what we do good at…and then AI doing what it’s good at, keeping things in sync,” Ritter said.

Ritter said there were two tests to see how AI could manage a system like a reactor on a smaller scale.

A heat pipe was used, and after the test, the AI was 0.3% off from a human operating it.

The other test was done with a digital five-watt reactor in a project done in conjunction with Idaho State University. The reactor was connected to the cloud and to an AI model to predict how the ISU students would use it.

“That was the world’s first digital twin limiting reactor (test) to happen here in the state,” Ritter said.

=htmlentities(get_the_title())?>%0D%0A%0D%0A=get_permalink()?>%0D%0A%0D%0A=htmlentities(‘For more stories like this one, be sure to visit https://www.eastidahonews.com/ for all of the latest news, community events and more.’)?>&subject=Check%20out%20this%20story%20from%20EastIdahoNews” class=”fa-stack jDialog”>





Source link

Share your love