Tony Grout, chief product and technology officer at M-Files, explains how organisations are unintentionally exposing sensitive business information through AI tools, as poor content governance becomes one of the biggest barriers to trusted AI.

Artificial intelligence has rapidly become a business priority, with organisations across every sector investing heavily in AI-powered assistants, automation tools and enterprise platforms. Yet despite growing adoption, many organisations are unknowingly increasing their exposure to information security and compliance risks.

According to M-Files, the problem isn’t the AI itself. Instead, businesses are layering sophisticated AI technologies on top of poorly organised and inconsistently governed information, allowing AI tools to surface content that may have been overshared long before AI was introduced.

Too many organisations still believe buying AI software makes them AI-ready. It doesn’t. Businesses have an information problem; AI simply exposes the weaknesses that already exist in how information is managed.

If content is poorly organised, inconsistently classified or governed by outdated permissions, AI amplifies those issues significantly. That affects productivity, undermines confidence in AI outputs and creates unnecessary compliance risks.

One of the biggest risks identified by the AI Readiness Model is what we describe as the “Copilot liability zone”. This occurs when organisations deploy AI over content that has been overshared through outdated folder permissions rather than robust document-level governance. While these permission issues often remain hidden during normal day-to-day operations, AI can expose them by surfacing sensitive information that users were unintentionally granted access to in the first place.

As organisations continue scaling enterprise AI across their operations, we have warned  that information governance must become a strategic priority if businesses are to realise AI’s full potential.

Every new AI deployment increases the importance of information governance because AI relied entirely on the quality of information it already has. Without strong metadata, content governance and permission controls, organisations are simply creating faster ways to generate unreliable answers instead of producing intelligent systems. As such, they risk undermining the very productivity gains AI is intended to deliver.

As AI becomes increasingly embedded within everyday business operations, M-Files believes organisations must shift the conversation beyond AI adoption and focus instead on whether the information underpinning those systems is accurate, governed and secure enough to support trusted AI at scale.

The organisations that will gain the greatest value from AI won’t necessarily be those deploying the most technology. They’ll be the ones that have invested first in trusted information. Information readiness is becoming the foundation of AI readiness, and that’s what will separate AI leaders from AI failures over the coming years.