Generative AI tools have revolutionized the way professionals interact with lengthy and complex documents, providing a new level of efficiency and insight. Ahmad Sultan, Privacy Officer and Cybersecurity Policy Manager, shared with us how he uses AI tools to streamline document review and enhance understanding.
As part of his weekly work at the Lab, Ahmad must review intricate government policies, regulations, and technical papers. These documents–often dense products written by legal and technical experts–are crucial to his work but can be time-intensive to process. Ahmad found that ChatGPT offers a significant advantage, allowing him to extract information relevant to technology policy considerations at Berkeley Lab and summarize key points efficiently.
Enhancing Document Review and Comprehension
For Ahmad, ChatGPT acts as a conversational assistant that simplifies his research and review of long and multi-layered documents. Whether analyzing government regulations or deciphering scientific terminology, the AI tool allows him to summarize complex content. This analysis allows him to prioritize which areas require deeper attention. AI also helps unpack any technical jargon he may encounter in his readings, breaking it down into terms that lend themselves more clearly to his work, enabling better communication and implementation of IT and Cybersecurity policy measures.
“Using ChatGPT for background reading is like switching from a bicycle to a car,” Ahmad explains. “It helps accelerate my understanding, and increases the volume of documents I can sift through without compromising the quality of my work.”
Best Practices for Cybersecurity Professionals
While the benefits are clear, Ahmad emphasizes the importance of caution when using AI tools. Sensitive information should never be input into ChatGPT or similar platforms unless explicitly approved. He identifies the following as categories to avoid sharing in unlicensed AI tools:
- Personal information: Any data that can identify individuals, such as names, addresses, or social security numbers
- Proprietary information: Confidential business details or trade secrets
- Controlled Unclassified Information (CUI): Sensitive government information that requires controlled access
- Non-Disclosure Agreement (NDA) content: Information protected by legal agreements
- Expert-controlled data: Information that falls under strict regulatory or institutional controls
Instead, Ahmad advocates using internal or locally licensed AI models for handling proprietary or sensitive data. This approach ensures compliance with legal and ethical standards while still leveraging the power of AI.
“While ChatGPT is powerful, it is not infallible,” cautions Ahmad. “Always verify and fact-check vigorously.”
in the bottom right or navigate to