Agentic/autonomous AI & runtime guardrails
Key Questions
What are key statistics on shadow AI in enterprises from FireTail?
FireTail reports 80% shadow AI usage with only 34% controls in place. This contributes to significant AI security risks including LLM and agentic threats.
What does the MIND report say about CISO confidence in AI security?
65% of CISOs doubt their data security controls for AI, despite 90% of organizations running GenAI at scale. This underscores gaps in enterprise AI security.
What is the GrafanaGhost vulnerability?
GrafanaGhost is a zero-click vulnerability enabling enterprise data exfiltration via indirect prompt injection and client-side bypasses. Attackers can force Grafana to leak sensitive data without user interaction.
How is Claude addressing code security?
Claude implements code sandboxing to isolate, constrain, and secure code in production. This mitigates remote code execution (RCE) risks in agentic AI.
What updates has OWASP made to GenAI security?
OWASP updated its GenAI Security Project with a new tools matrix. It covers risks and defenses for generative AI applications.
FireTail 80% shadow AI/34% controls; MIND 65% CISO doubt; GrafanaGhost zero-click vuln; OpenClaw growth/CVE; Claude RCE/sandboxing; OWASP GenAI update; OpenAI patches; NemoClaw/Agent-Infra; Mimecast Incydr; NDSS/Mythos ZT; OpenShift CoCo; Cisco/Databricks/Fortinet/Oasis/Bonfy; NVIDIA/RSAC.