The Outsized Shadow: Why 5% of AI Users Are Your Biggest Security Risk

Big security risks come in small packages. While enterprise security teams focus on policing the proliferation of employees using ChatGPT and Claude for quick drafting tasks, a more urgent threat is posed by a handful of AI super-adopters who are quietly hardcoding unvetted tools into critical business operations. According to new research published by Akamai, the top 5% of enterprise power

The article reports that new Akamai research finds the top 5% of AI “power users” in enterprises are quietly hardcoding unvetted AI tools into critical business operations, creating a disproportionate security risk relative to casual users who rely on services like ChatGPT and Claude for lightweight drafting tasks. These super-adopters integrate external or shadow AI/SaaS tools directly into workflows and systems without formal vetting, change control, or security review, increasing the likelihood of data exposure, dependency risk, and operational disruption. From RealGround’s perspective, this pattern represents a concentrated SaaS AI risk cluster that requires identifying and inventorying unsanctioned AI integrations, establishing governance around which AI tools can be embedded in business processes, and instituting controls for approval, monitoring, and decommissioning of AI-based SaaS dependencies. Practically, organizations should conduct readiness assessments to map high-risk AI usage, enforce policy-based guardrails for AI tool adoption, and implement continuous oversight for departments and roles most likely to become AI super-adopters.