At the RSA Conference 2025 in San Francisco, Cisco Systems Inc. introduced new AI-driven security innovations aimed at helping enterprises tackle AI-related threats and enhance their overall security measures.

Cisco also upgraded its Industrial Threat Defense solution, integrating tools like Cisco Vulnerability Management and Splunk OT Security to bolster protection for industrial networks through better visibility and risk prioritization.

Skyhigh Security enhanced its SSE platform with Data Security Posture Management capabilities, improving data visibility and compliance for organizations.

Among the innovations, Cisco's Automated XDR Forensics capabilities significantly improve visibility into endpoint activities, which enhances the accuracy of investigations and response times for security operations.

Research presented by Mindgard revealed vulnerabilities in AI guardrails that can be exploited using techniques such as emoji and Unicode smuggling, underscoring the need for stronger protective measures.

In related findings, Backslash Security's research indicated that leading large language models, particularly GPT-4o, generate vulnerable code by default, prompting the introduction of new protective capabilities.

X-PHY debuted a real-time deepfake detection tool that operates offline, analyzing various media for manipulations to protect organizations from fraud and misinformation.

BreachLock introduced Adversarial Exposure Validation (AEV), a tool that allows security teams to simulate real threat actor exploits and prioritize vulnerabilities effectively.

Delinea upgraded its identity security platform to include features that protect AI systems, such as shadow AI detection and least privilege controls.

Aryon Security launched a Cloud Security Enforcement Platform aimed at minimizing cloud risks and human error, emerging from stealth mode with $9 million in funding.

GTB's 'Get the Best™' approach emphasizes top-tier protection for intellectual property and sensitive data in the evolving landscape of AI.

Wallarm unveiled Agentic AI Protection, designed to secure AI agents against various attack vectors, thereby enhancing API security.