According to Checkpoint, 87% of organisations cannot enforce security policies on traffic to AI SaaS services in real time and 89% of organisations say their endpoint tools cannot detect and control unauthorised AI applications or browser-based AI tools on managed devices (1). Add to this that 68% of organisations do not even have policies in place to prevent shadow AI (2) and you have a bleak picture of the current environment.

87%

of organisations cannot enforce security policies on traffic to AI SaaS services in real time.

ig-01 robot reading up on processes-on-tablet

It is clear, we live in a world of ever-increasing AI usage where visibility and policy are not keeping pace.

The challenge is complicated further by ever-increasing cybersecurity regulation such as the EU AI Act and NIST AI RMF, and the mounting pressure on leaders to act. But who owns this risk and who should act? Checkpoint’s research indicates that ownership is split: 44% with the CISOs, 40% with cross-functional committees and 36% with the CIO or IT Leadership. Furthermore only 14% have a dedicated AI Security leader (1). With this fragmented approach to accountability, enforcing your policies, however robust they are, is going to prove challenging.

Accountability can clearly be identified as a key challenge; while most organisations do not have robust policies or enforcement mechanisms in place this is likely due to a lack of clear ownership and accountability at the board level. Organisations must act with haste to clearly identify an AI Security leadership structure before establishing policy and governance, if they want to ensure secure and responsible AI adoption.

Clear leadership is a key element in your path to secure AI, but it needs to be supported through effective policy enforcement and oversight. This takes us back to two of Infinigate’s key areas of AI Security: AI Powered and AI Platforms. A clear leadership structure and an established strategy are essential requirements before you invest in the right security tools, AI guardrails and platforms to make secure AI an achievable reality. To get access to industry leading advice and guidance on what this looks like in practice, please reach out to your local Infinigate representative who will be happy to help you.

Want all five pillars now?

Get access to industry leading advice and guidance on what this looks like in practice.

Dan Walker

Group Business Development Director

  1. Checkpoint’s 2026 Cloud Security Report, Securing the AI Transformation (Cloud Security Report 2026 | Check Point Software)
  2. IBM’s Cost of a Data Breach Report 2026 (Cost of a Data Breach Report 2026 | IBM)