The new artificial intelligence (AI) model from China called DeepSeek created a stock market meltdown on Monday, with the Nasdaq composite dropping 3% and the S&P 500 falling 1.5%. Beyond hammering the share prices of the world’s most valuable companies, DeepSeek has potential implications on vast swaths of America’s innovation industries—including energy.
COMMENTARY
While U.S. technology companies must quickly respond to the challenges posed by the new DeepSeek model, and the AI innovations to come, other businesses—like the energy companies currently exploring uses for AI in their operations—have a different responsibility. Utilities, independent power producers, and energy companies of all stripes, must take a more measured approach and use this as a teachable moment for their employees to understand the safety and security risks inherent in AI tools. They have to underscore that employees should treat new AI tools no differently from other technologies that enter the enterprise, and use the safety and security standards that inform every decision on technology adoption.
Artificial intelligence has the incredible potential to make energy facilities—and particularly nuclear energy facilities—easier to develop, operate, orchestrate, and maintain. But only if these applications can adhere to the strictest standards of data security, privacy, and operational integrity. Nowhere is this more important than among the nation’s nuclear fleet operators.
Understanding the Risks of DeepSeek R1
DeepSeek R1, an AI model developed by the Chinese company DeepSeek, has remarkably impressive reasoning, problem-solving, and coding capabilities at a fraction of the cost and with a fraction of the energy requirements of its competitors. Its open-source nature makes it accessible to anyone and its launch will undoubtedly be recognized as a Cambrian moment for AI across the global economy. However, for critical sectors like energy (and particularly nuclear energy) the risks of racing to adopt the “latest and greatest AI” models outweigh any potential advantages.
Open-source AI models have significant benefits thanks to their transparency and ability to foster collaboration. Yet, with this openness comes a need for diligence, especially when a model originates from a country like China, where data handling and security practices differ from those in the U.S. or other regions.
DeepSeek R1’s rapid adoption highlights its utility, but it also raises important questions about how data is handled and whether there are risks of unintended information exposure. For example, AI models often learn from the data they process. So, any sensitive company information—from plant operations data, maintenance records, or security protocols—could become part of the model’s learning process if the tool is improperly used. Risks like these are especially concerning in the nuclear sector, where safeguarding critical infrastructure is paramount.
Practical Steps for the Nuclear Industry
To mitigate risks, nuclear power plants must act proactively to ensure that tools like DeepSeek R1 are used responsibly. Here’s how:
- Educate Employees. Employees must understand the importance of safeguarding sensitive information when using AI tools. Plant operators should issue clear guidance advising against using DeepSeek R1 for work-related tasks or sharing company data with it.
- Clarify Policies on AI Usage. Organizations should establish policies that define how and when AI tools can be used. These policies should emphasize the importance of using vetted and approved models to ensure security.
- Evaluate AI Models Thoroughly. While DeepSeek R1 offers a version that can be hosted internally, any implementation should undergo a rigorous review process to verify that it meets security and compliance standards.
- Collaborate to Set Standards. Industry-wide collaboration is essential to create best practices for evaluating AI tools in critical infrastructure. Organizations like the Nuclear Energy Institute (NEI) can play a pivotal role in guiding how AI is integrated into the industry.
It’s important to note that open source itself is not the issue—far from it. Open-source innovation has driven countless advancements in technology, including in the nuclear sector. However, the open nature of tools like DeepSeek R1 means they can be accessed by anyone, including those who might misuse them or exploit vulnerabilities. This dual-edged nature of open-source AI is why thoughtful evaluation and cautious implementation are necessary, particularly for industries with heightened security requirements.
A Balanced Approach to AI Adoption
AI tools like DeepSeek R1 represent a remarkable leap forward in what technology can achieve. They have the potential to improve efficiency and decision-making across many industries. However, for sectors like nuclear power, where security is non-negotiable, it is critical to approach such tools with care.
The goal is not to reject innovation but to embrace it responsibly. By educating employees, implementing clear policies, and thoroughly evaluating new tools, we can ensure that AI contributes to the safety and success of the nuclear industry without introducing unnecessary risks.
Nuclear power is the foundation of a clean energy future. Let’s continue to protect it by using advanced technologies responsibly and with the vigilance it deserves.
—Trey Lauderdale is CEO of Atomic Canyon, whose advanced AI-powered solution is transforming nuclear data management, beginning with Neutron—a solution that navigates billions of pages of technical documentation, enhances Nuclear Regulatory Commission (NRC) data access, and unlocks workflow efficiency.
