# Data Protection Methods with LLMs How to keep company data safe with LLMs? - **Source:** [Software Engineering Radio - The Podcast for Professional Software Developers](https://brifing.si/source/45a0c92c-2388-493f-9768-fa13bb6ed2f8) - **Published:** 2026-08-19 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/0518f50f-b0ae-449c-b20b-0ad3e0a906f1 ## Data Security Issues with LLMs In enterprise AI deployments, LLMs often reside outside the cloud tenant. This means that with every query, RAG retrieval, and tool call, data leaves the company's perimeter. This can lead to data being used by LLM providers for training and audits. While contracts can restrict this usage, they do not prevent prompt injection or unintended data exposure. ## Data Leakage Risks Company data is transferred to LLMs through natural language queries, APIs, tool and function calls. This process carries the risk of data being used or misused by LLM providers. Attacks like prompt injection pose a serious threat to data security. ## Security Measures Sathiesh Veera emphasizes that security measures and data filtering methods can be employed at each layer to comply with data security policies. These measures aim to protect data as it is transferred to LLMs and prevent unauthorized access. ## Data Filtering Data filtering ensures that the data sent to LLMs does not contain sensitive information. This is a critical step for protecting financial, personal, or trade secrets. Filtering mechanisms clean data before it is processed by LLMs. ## Layered Security Approach A layered security approach involves implementing data protection strategies at every stage. This ensures that different security measures are activated during data input, processing, and output. This approach minimizes the risk of data leakage. ## Auditing LLMs Regular audits should be conducted during the use of LLMs. These audits evaluate how data is used and the effectiveness of security measures. Audits help identify potential security vulnerabilities. ## Training and Awareness It is important to train employees on the use of LLMs and data security. This helps prevent behaviors that could inadvertently lead to data leaks. Awareness programs strengthen the security culture. ## Future Challenges As LLMs become more widespread, data security challenges will increase. Companies must develop continuously updated security strategies to address these challenges. In the future, more complex security measures and methods will be required. ## Sources covering this story - https://se-radio.net/2026/08/se-radio-734-sathiesh-veera-on-engineering-data-protection-guardrails-with-llms/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._