Concentric AI announced a DSPM solution with support for optimized large language models that provide enhanced data security and privacy.
As a result, Concentric AI’s semantic intelligence delivers a semantic understanding of data and leverages context to provide pinpoint accuracy in detecting sensitive data ranging from intellectual property to financial data to sensitive business information, PII/PCI/PHI and more.
According to Gartner, “Foundation models (which are mostly large language models) are destined to replace task-specific models”… “Foundation models represent a major shift in the field of artificial intelligence due to their massive pre-training and significant improvements in accuracy over one variety of tasks.”
With massive data volume growth, data proliferation in cloud and on-premises data storage, and the tremendous complexity inherent in the nature of sensitive data, organizations often struggle with the need for large teams and huge operational costs to deal with massive data security risk burdens Security teams writing rules and matching patterns to discover sensitive data. This can leave organizations with limited visibility into the location of business-critical data and an inability to identify risks related to incorrect permissions, inappropriate permissions, risky disclosure and unauthorized access.
Unlike other solutions, Concentric AI uses optimized large language models to accurately recognize data without the need for imprecise regex or pattern matching. While large language models are very data intensive, highly optimized data compression models from Adaptive Manifold Compression achieve up to 10x compression and orders of magnitude improvements in data recognition rates for optimal computational efficiency, storage, and performance.
With Concentric AI, organizations discover all sensitive data with context, accuracy, and minimal false positives and negatives, all with unmatched speed and efficiency. Concentric AI’s sophisticated AI-based technology uniquely performs computations on large language models without the computational burden or inefficiency of other solutions.
“The popularity of models like GPT-3 has attracted the attention of CIOs, CTOs, CISOs, and data and analytics leaders looking to realize their potential for business use cases,” said Karthik Krishnan, CEO of Concentric AI.
“As the only company leveraging large language models, Concentric AI has the advantage of enabling organizations to protect their data with unmatched contextual understanding, all in an accurate and efficient manner required to handle the complex and ever-growing Tackle today’s data environments and the security risks that come with them. These models enable organizations to go beyond regex or pattern-based detection of sensitive data to understand semantic meaning with unmatched efficiency while maintaining high accuracy and minimal false positives and false negatives,” concluded Krishnan.
Concentric AI’s semantic intelligence enables fingerprinting of AI data to avoid rescanning data to look for new patterns. This allows organizations to discover patterns and specific things they are looking for without the burden of repetitive scanning. Once scanned, the data is fingerprinted to discover features that will be important in the future.
According to recent industry studies, the average cost of a data breach in 2022 was $4.35 million, hitting an all-time high. Breaches in organizations with fully deployed security AI and automation cost $3.05 million less than breaches in organizations without security AI and automation deployed.
Finally, the research showed that organizations with fully deployed security AI and automation had an average of 74 days less time to identify and contain the security breach, known as the breach lifecycle, than organizations without security AI and automation.
Concentric’s new capabilities extend the company’s AI technology to support highly differentiated, powerful capabilities for effective data security management. Zero training data models enable data to be automatically categorized without requiring customers to provide training datasets.
The solution includes hundreds of ready-to-use models to categorize data from tax returns to source code to contracts, trade data, PHI/PII/PCI and more, allowing customers to operationalize data security without much upfront or ongoing work. or require large teams.
As a result, to date, Concentric AI has protected more than 1 million users from security breaches, scaled data per customer to petabytes with unmatched accuracy, and remediated and protected up to 20 percent of mission-critical data per customer for erroneously over-distributed permissions, risky sharing, inappropriate Permissions or Unauthorized Access.
Concentric AI’s DSPM solution scans companies’ data, detects sensitive or mission-critical content, identifies the most appropriate classification category, and automatically flags the data. Concentric AI uses artificial intelligence (AI) to improve detection and classification accuracy and efficiency to avoid endless regex rules and inaccurate end-user labeling.
In addition, Concentric AI can monitor and autonomously identify risks to financial and other data from improper approvals, incorrect permissions, risky sharing, and unauthorized access. It can automatically fix permissions and sharing issues, or leverage other security solutions and cloud APIs to quickly and continuously protect exposed data.
Concentric AI’s semantic intelligence automates the security of unstructured and structured data using deep learning to categorize data, uncover business criticality, and reduce risk. Its risk distance analysis technology uses the basic security practices observed for each category of data to detect security anomalies in individual files.
It compares documents of the same type to identify risks of over-approval, third-party access, mislocation, or misclassification. Organizations benefit from the expertise of content owners without intrusive classification rules, without the need for rules, regex, or policy maintenance.
Concentric AI’s latest version of Semantic Intelligence with optimized large language models is now available.
