Concentric AI’s DSPM finds sensitive or business-critical material 2023

Concentric AI has released a DSPM solution that supports big language models that are designed for data confidentiality and protection.

Concentric AI’s Semantic Intelligence provides semantic comprehension of data and utilizes context to enable exact accuracy in detecting sensitive material, including intellectual property, financial data, company private information, PII/PCI/PHI, etc.

Foundation models, which are mostly big language models, are intended to replace task-specific models, according Gartner.

Foundation models represent a significant advancement in the area of artificial intelligence because to their extensive pretraining and strong accuracy increases across a wide range of tasks.

Faced with massive data volume growth, data sprawl across cloud and on-premises data repositories, and massive complexity in the nature of sensitive data, organizations have struggled with requiring large teams and massive operational costs to address massive data security risks, frequently burdening security teams with rule writing and pattern matching to discover sensitive data. This can lead to enterprises having limited insight into the location of mission-critical data and an inability to recognize hazards associated with incorrect entitlements, improper permissions, unsafe sharing, and illegal access.

Concentric AI, unlike other systems, leverages optimized huge language models to discover data precisely without the need for erroneous regex or pattern matching. Despite the fact that big language models are extremely data-intensive, highly efficient data compression Adaptive Manifold Compression models deliver up to tenfold compression and orders of magnitude gains in data discovery rates for best compute efficiency, storage, and performance.

With Concentric AI, businesses are able to uncover all critical data with context, precision, and minimum false positives and negatives, all at an unrivaled rate and efficiency. Concentric AI’s distinct AI-based technology conducts calculations on massive language models without the computational burden or inefficiency of competing alternatives.

“CIOs, CTOs, CISOs, and data and analytics professionals are interested in exploiting the potential of models like as GPT-3 for commercial use cases,” said Karthik Krishnan, CEO of Concentric AI.

“Being the only company to leverage large language models gives Concentric AI the advantage of enabling organizations to protect their data with unmatched contextual understanding, all with the accuracy and efficiency required to address today’s complex and ever-expanding data environments and the associated security risks. “Our models enable organizations to transcend beyond regex or pattern-based detection of sensitive data in order to comprehend semantic meaning with unrivaled accuracy and minimum false positives and negatives,” Krishnan said.

Concentric AI’s Semantic Intelligence enables data fingerprinting to prevent rescanning for new patterns. Hence, enterprises are able to uncover trends and particular items without the strain of recurrent scanning. After being scanned, the data is fingerprinted in order to find future-relevant traits.

According to current industry studies, the average cost of a data breach reached a record high of $4.35 million in 2022. Breach costs for firms with fully implemented security AI and automation are $3.05 million less than for organizations with no deployed security AI and automation.

The report also revealed that firms with fully integrated security AI and automation witnessed an average 74-day reduction in the breach lifecycle, which is the time required to discover and contain a breach.

Concentric’s new capabilities broaden the company’s AI technology to enable highly distinctive, potent data security posture monitoring capabilities. Zero training data models provide automatic data classification without requiring clients to furnish training datasets.

The system includes hundreds of out-of-the-box models for categorizing data, such as tax filings, source code, contracts, trading data, PHI/PII/PCI, and more, allowing clients to operationalize data protection without extensive upfront or ongoing labor or huge teams.

As a result, Concentric AI has protected more than one million users from data breaches, scaled to petabytes of data per customer with unmatched accuracy, and remedied and protected up to twenty percent of business-critical data per customer that has been overshared due to incorrect entitlements, risky sharing, inappropriate permissions, or unauthorized access.

Concentric AI’s DSPM solution examines an organization’s data, discovers sensitive or business-critical material, determines the best applicable categorization category, and tags the data automatically. Concentric AI utilizes artificial intelligence (AI) to enhance finding and categorization accuracy and efficiency, hence avoiding lengthy regex rules and erroneous end-user labeling.

In addition, Concentric AI is able to monitor and automatically identify threats to financial and other data posed by improper permissions, incorrect entitlements, unsafe sharing, and illegal access. It can automatically fix permissions and sharing problems, as well as utilize other security solutions and cloud APIs to swiftly and continually safeguard exposed data.

Concentric AI’s Semantic Intelligence automates the security of unstructured and structured data using deep learning to classify data, identify business-criticality, and decrease risk. Its Risk Distance analysis tool employs the baseline security practices observed for each data category to identify security abnormalities in specific files.

It examines documents of the same kind in order to discover risks associated with oversharing, third-party access, incorrect location, and misclassification. Companies gain from the knowledge of content owners without onerous categorization regulations and without the need to maintain rules, standards, or policies.

The most recent version of Concentric AI’s Semantic Intelligence with improved big language models is now available.

Leave a Comment