Data Anonymization (Secure De-Identification Inference: SDI)
I. Applicable Users
Users of AskTable's private deployment who utilize cloud computing power (large models).
II. Function Overview
When using AI technology for enterprise data retrieval and analysis, privacy and security have become key concerns for enterprises.
To address this, AskTable introduces the SDI (Secure De-Identification Inference) anonymization feature, which is used to perform local encryption and anonymization of sensitive information.
This feature supports anonymizing sensitive fields (such as names, phone numbers, ID numbers, email addresses, bank card numbers, etc.) locally, ensuring that the identifiers are removed before the data is sent to the cloud AI model for inference, thus preventing the leakage of sensitive information.
SDI realizes the concept of "data stays local, thinking happens in the cloud," ensuring data security while fully utilizing the intelligence of high-end cloud-based large models.
III. Supported Anonymized Field Types
AskTable currently supports the following types of fields for anonymization:
| Type | Original Data | Fake Data |
|---|---|---|
| Name | Zhang Fei Yi | Wang Ming |
| Phone Number | 13812341234 | 13900000001 |
| ID Number | 110101199001011234 | 310104199208150001 |
| Email Address | 123456789@qq.com | c000001@example.com |
If you have any other requirements, please contact us.
IV. Configuration Method and Usage Process
AskTable supports setting up anonymization policies at the data source field level. With system configuration, the anonymization function can be enabled without additional development or complex deployment.
Configuration steps:
- Enter the data source field management interface;
- In the "Anonymization Encryption" column, select the type of sensitive information for the corresponding field (e.g., "Name", "Phone");
- After saving, the field will automatically be anonymized in the query results.
As shown in the figure below, you can configure the anonymization types for the salesperson_name and phone fields respectively:
V. Frequently Asked Questions
1. Does anonymization affect the accuracy of analysis?
In most cases, it does not. When generating SQL queries, AskTable still uses real field names and structures. It only anonymizes sensitive fields during the query result phase, which does not affect the semantic meaning or accuracy of the analysis. However, in a few rare tasks, noise introduced during the restoration process may lead to misunderstandings by the large model, resulting in phenomena such as "unable to answer."
2. Will enabling SDI affect performance?
There will be some performance overhead. For queries involving anonymized fields, the average processing time increases by approximately 2–3 seconds.
3. Is it necessary to enable the SDI feature in the SaaS version?
It is not recommended to enable it. The SaaS version runs in a cloud environment, and data is already transmitted to the AskTable backend for processing by default. Enabling SDI would have limited significance and could add extra latency.
4. Does enabling SDI affect business users?
No impact. SDI processing is completed internally within the system, and business users are unaware of it. It also does not affect their experience when asking questions, querying, or viewing results.
VI. Summary
The SDI (de-identification) feature is an important capability of AskTable that balances data security with intelligent analysis. By anonymizing sensitive fields locally and then sending the processed data to the cloud model for inference, SDI realizes a new model of "data remains local, intelligence relies on the cloud."
Compared to private deployment of large models, the cost of using cloud computing power is only about 1% of that. The introduction of SDI enables enterprises to access the most advanced AI capabilities at a low cost and securely, significantly reducing the entry barriers.
For industries with high data security requirements and sensitivity to costs, such as government and enterprise sectors, healthcare, etc., SDI provides a solution that balances privacy protection and intelligent analysis. This helps accelerate the practical application of AI technology in real business scenarios and promotes the popularization and development of enterprise intelligent transformation.
Through SDI, AskTable not only ensures data security but also provides users with cost-effective intelligent analysis capabilities, truly achieving the AI value cycle of "secure and trustworthy, cost-controlled, and effective."