Cybanetix has introduced a Managed AI Service aimed at addressing AI security challenges within enterprises. The service covers employees' use of AI, governance, and embedded AI systems.
The service combines technologies from NOMA, SentinelOne, Microsoft, and Exabeam with Cybanetix's consultancy services and 24/7 monitoring capabilities. According to Cybanetix, the service can respond to alerts within 15 minutes and is designed to provide a unified approach to AI security.
As AI adoption increases across organisations, security controls may not always develop at the same pace. This can affect three primary areas:
- User Behaviour: Employees may use unauthorised AI models or upload data to public language models, creating potential security and compliance risks.
- Governance: Organisations may face challenges related to model provenance, AI bills of materials (AI-BOMs), and ownership and accountability for AI assets.
- Embedded AI: AI tools integrated into business processes may have elevated privileges that could increase security risks if not properly managed.
The Managed AI Service is intended to provide visibility across the AI environment. It incorporates observability, behavioural monitoring, infrastructure and application protection, and testing capabilities to help identify and address potential threats.
The service includes the following technologies:
- SentinelOne and Microsoft Purview for user-level controls.
- NOMA for AI discovery, access control, and response planning.
- Exabeam for agent behaviour analytics.
These technologies are supported by Cybanetix's consultancy services and security operations. The company also provides AI security policy development, policy implementation, and system hardening services.
The Managed AI Service includes:
- AI Risk Assessment: Inventory and risk mapping of AI-related assets and components.
- Technology Deployment: Deployment of observability tools and runtime protections aligned with AI usage policies.
- AI Posture Management: Integration with security operations centre (SOC) playbooks and policy enforcement within AI environments.
- Managed Detection and Response (MDR): Integration with SOC and MDR functions to support risk management activities.
- Continuous AI Risk Reporting: Ongoing reporting designed to support visibility and compliance requirements.
- Testing and Validation: Real-time testing to identify vulnerabilities and assess security controls.
The service is intended to support organisations in identifying AI-related risks, monitoring AI deployments, and managing security controls across AI environments.
As AI tools become more widely used across business functions, organisations may require additional expertise and operational capabilities to address associated security risks. The Managed AI Service is designed to provide AI security support through a combination of technology, monitoring, and consultancy services.
