MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641107076 A) filed by Abburu Sai Shivaprakash on September 05, 2026, for System And Method For Cross-Plane Cost-Signal Security Monitoring Of Artificial Intelligence Infrastructure With Non-Human Identity Attribution, Unified Event Graph Correlation, And Immutable Billing Ledger Forensics.
Inventor includes Abburu Sai Shivaprakash.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: The present invention discloses the development of a system and method for security monitoring of artificial intelligence infrastructure utilizes cost-plane telemetry from billing systems as a primary security signal. It integrates token consumption metrics, model invocation records, and spend data with authentication and identity lifecycle records. A non-human identity attribution layer resolves cost records to identities through a hierarchy of mapping and validation processes, classifying unattributed records as such. A unified event graph engine constructs a real-time directed acyclic graph for event correlation and anomaly score propagation. Additionally, a cost-behavioral baseline engine creates per-identity baselines, while a composite threat scoring engine combines deviations into a score classified against a threat taxonomy, prompting specific security responses. An immutable billing ledger validates records against internal logs to identify integrity violations. Notable technical effects include rapid score generation, high attribution completeness, and the detection of insider threats and AI-specific attacks unidentifiable by traditional monitoring systems. FIG.1
Disclaimer: Curated by HT Syndication.
