System and method for preventing identity theft or misuse by restricting access

What is this patent about?

’374 is related to the field of data security and network file management. It specifically addresses the challenges faced by large enterprises, such as financial institutions, in identifying and protecting personally identifiable information (PII) scattered across vast numbers of unstructured data files. The background context involves the increasing legal and regulatory pressure to prevent identity theft by securing sensitive data like credit card numbers and social security numbers stored in archived emails or scanned documents.

The underlying idea behind ’374 is that sensitive data can be accurately identified without manual review by using a two-stage statistical and algorithmic filter. Instead of simply searching for keywords, the system evaluates the pattern density of potential sensitive strings relative to the file size and then applies a secondary validation layer. This secondary layer uses the mathematical properties of the data itself—specifically the presence of a valid check digit—to distinguish between random alphanumeric noise and actual sensitive records.

The claims of ’374 focus on a computer-implemented method and system that scans files for specific alphanumeric strings, calculates the density of these strings, and performs a Luhn algorithm or similar check-digit validation on the identified data. Once a file is validated as containing sensitive information, the claims cover a specific suite of restrictive actions, including the use of privilege masks, crypt checksums to prevent covert code execution, and the dynamic granting and revoking of identifiers during the file-opening process.

In practice, the invention operates by first isolating files where the frequency of account-like numbers suggests a high probability of sensitive content. By dividing the number of occurrences by the file size, the system avoids over-flagging large files that contain only incidental matches. The subsequent validation of the check digit ensures that the system only restricts access to files containing mathematically valid account numbers, which significantly reduces the false alarm rate compared to traditional keyword searches.

This approach differs from prior solutions by integrating the detection logic directly with a multi-faceted enforcement engine. Rather than just flagging a file for a later administrator response, the system can automatically implement site-specific commands to gather evidence of unauthorized access attempts or apply encryption and password protection in real-time. By focusing on the statistical likelihood of genuine data through pattern density and mathematical verification, the invention provides a scalable way to secure data without moving it to a centralized database.