The Department of Energy Releases Comprehensive Guide for Generative AI Use

You can read the full article here.

The Department of Energy (DOE) has introduced a 61-page reference guide for using generative AI tools, such as ChatGPT, by its employees and contractors. Published on June 14 and available on the DOE’s internal network, this guide aims to educate DOE personnel on the responsible use of AI technology, outlining best practices, potential use cases, and existing government resources related to AI. Although not formal policy, the document emphasizes privacy and security, advising against inputting nonpublic data into AI systems unless appropriate measures are in place. The guide recommends conducting privacy impact assessments, tailored AI training for employees, and lifecycle data stewardship.

The guide, developed by DOE’s generative AI (GenAI) Tiger Team, also discourages using GenAI to create confidential or mission-critical information. It also highlights the importance of understanding data rights in contract solicitations and encourages prompt engineering training to mitigate AI bias. The guide is a living document designed to evolve with technological and regulatory advancements, reflecting the DOE’s commitment to leveraging AI responsibly and effectively.

Why Read This Article?

Insight into DOE’s AI Integration: Understand how a major federal agency is approaching the integration of generative AI, providing a potential model for other agencies.

Best Practices for AI Use: Gain knowledge of best practices for generative AI use, emphasizing privacy and security, relevant to public sector AI deployment.

Risk Mitigation Strategies: Discover strategies for mitigating AI risks, such as privacy impact assessments and prompt engineering training.