Check Point Research's AI Security Report 2026 reveals that AI has evolved from a preparatory tool to an active operator in cyber attacks, executing live intrusions across diverse threat actors from nation-states to cybercriminals. AI-generated malware and attack suites are now deployment-ready and often conceal AI involvement. Attackers exploit commercial AI models with persistent bypass configurations, and AI-enabled criminal services like phishing-as-a-service and conversational AI voice agents have matured. The report also notes the increasing sophistication of multi-channel social engineering using forged biometric and identity data. The AI ecosystem itself introduces new vulnerabilities, including indirect prompt injection attacks and supply-chain risks. Additionally, enterprise data leakage through generative AI is a growing concern, with high-risk prompts increasing and uneven risk distribution across sectors.

AI is now directly involved in executing cyber attacks, increasing the speed, scale, and sophistication of intrusions. The use of AI-generated malware and attack frameworks lowers the barrier for attackers and expands the threat landscape. AI-enabled criminal tooling markets facilitate large-scale phishing and vishing campaigns. The forgery of biometric and identity data enhances social engineering effectiveness. The AI models and their ecosystems introduce new vulnerabilities and supply-chain risks. Enterprises face persistent and growing risks of sensitive data leakage through generative AI usage, with some sectors more affected than others.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Organizations should monitor updates from Check Point Research and other authoritative sources for specific mitigations related to AI-enabled threats. Given the evolving nature of AI threats, security teams should prioritize risk assessment of AI usage within their environments and apply vendor-recommended controls as they become available. No vendor advisory indicates that no action is currently required or that the issue is already mitigated.