Three related pending U.S. patent families.
Continuous Trust-Signal-Based Constraint of AI Agent Authorization Through Artifact Mutation and Velocity-Triggered Enforcement
Delegated authority and runtime enforcement for agents acting on behalf of human principals.
- delegated AI-agent authority
- continuous authorization
- authorization-artifact mutation & revocation
- permission-tier adjustment
- multi-agent delegation
- pre-execution policy enforcement
- fail-secure degradation
- rate-of-change security triggers
- cryptographically attributable audit records
An agent's effective authority stays bound to the current verified state of the originating human or organizational authority, rather than remaining fixed at the time of delegation.
Neural Intelligence System for Continuous Biometric Authentication Within Zero Trust Security Architecture
Reassessing identity assurance continuously across an active session.
- continuous session authentication
- risk-adaptive thresholds
- on-device biometric processing
- derived biometric representations
- dynamic privilege adjustment
- least-privilege enforcement
- liveness / presentation-attack detection
- step-up authentication
- session-bound identity assurance
Neural Intelligence System for Social Networking Applications, Multi-Platform Biometric Security, and AI-Powered Deception Detection
Combining signals across modalities to assess identity and deception risk.
- multimodal biometric analysis
- facial geometry
- behavioral patterns
- cross-platform identity consistency
- NLP-based deception indicators
- metadata / image analysis
- aggregate risk scoring
- adaptive security responses
Strong alignment with major security frameworks.
Provisional filings
from August 2025
The portfolio addresses security problems now central to U.S. and allied cybersecurity frameworks.
Its focus — continuous authentication, dynamic authorization, least privilege, delegated authority, runtime verification, cryptographic accountability, revocation, and fail-secure control of autonomous systems — maps onto the Department of Defense Zero Trust architecture and NSA's 2026 implementation guidance, NIST's AI Agent Identity and Authorization initiative, the joint Five Eyes agentic-AI guidance, and subsequent NSA AI Security Center guidance.
The research program traces to provisional filings beginning August 19, 2025. Portions of the portfolio were filed before several of these 2026 publications — including the April 2026 Five Eyes guidance and the May 2026 NSA AI-security guidance, with the parent application predating the January 2026 NSA Zero Trust implementation guidelines.
- 2022
- U.S. DoD — Zero Trust Strategy Source
- Jan 30, 2026
- NSA — Zero Trust Implementation Guidelines, Phase One & Two Source
- Feb 5, 2026
- NIST — Software & AI Agent Identity and Authorization (concept paper) Source
- Apr 30, 2026
- Five Eyes — Careful Adoption of Agentic AI Services (NSA · CISA · ACSC · CCCS · NCSC-NZ · NCSC-UK) Source
- May 2026
- NSA AISC — Security Design Considerations for AI-Driven Automation (MCP) Source
These references reflect independent convergence in the broader security field. They do not assert government endorsement, adoption, infringement, or that any external framework practices a Northgate patent claim.
A note on scope
These applications should not be interpreted as claims that any particular third-party product, government framework, or implementation practices the patent claims. References to external developments are provided only to examine the evolution of the broader security problem space.