Nearly half of enterprise AI use bypasses security, creating major risks from unmanaged 'shadow AI'.

Akamai's 2026 report reveals that nearly 50% of enterprise AI usage occurs outside corporate security controls, creating significant 'shadow AI' risks. A small group of power users drive most AI interactions, while many unmanaged AI apps and browser extensions expose companies to new cyber threats. Novel attack methods like Vibe hacking and CursorJacking bypass traditional defenses, targeting AI coding assistants and autonomous agents. Akamai advises firms to monitor high-risk users, enforce single sign-on, inspect AI interactions in real-time, vet extensions, and secure AI agents to mitigate these emerging risks.