Skip to content
IT
Cybersecurity · VIDEO

Shadow AI: how SaaS updates hide invisible risks in companies

20/11/2026 · 10:12 · GoYou.it Videos
Read the article Italiano

The IBM Cost of a Data Breach Report 2026 reveals that incidents related to shadow AI have doubled, but the main problem is not employees using unauthorized chatbots. The real risk lies in AI models integrated into SaaS and internal applications through vendor updates, often undetected by security teams. Let's explore how to build an inventory to identify these hidden vulnerabilities.

KEY TAKEAWAYS

  • Incidents related to shadow AI have doubled from 20% to 43% according to the IBM Cost of a Data Breach Report 2026
  • Shadow AI is not just about unauthorized chatbots: it includes models integrated into SaaS and internal applications through vendor updates
  • Most of these AI integrations happen transparently, through release notes ignored by security teams
  • A complete inventory of AI assets is essential to govern specific risks such as uncontrolled data flows to external models
  • Tools like ShadowAI-Watch can help identify undetected LLM connections in local processes

Transcript

**Radio News Bulletin:**

A recent analysis by Thales Imperva reveals that 86% of organizations have not correctly inventoried AI models integrated into existing software, creating significant security and compliance risks.

The term "Shadow AI" refers to two main issues: unauthorized chatbots used by employees and AI functionalities invisibly integrated into existing software through vendor updates. Alarmingly, 43% of security incidents now involve non-inventoried AI, with most breaches occurring through SaaS software updates.

This visibility gap exposes companies to data breach risks and non-compliance with regulations such as the NIS2 Directive and the DORA Regulation. Unmanaged AI integrations create uncontrolled paths between sensitive data and third-party infrastructures, posing unique security challenges.

To mitigate these risks, organizations must first discover all active AI models within their infrastructure. This discovery process is now the top priority for security teams, as outlined in the Thales Imperva AI Application Security Framework.

A complete AI inventory should include customer-facing applications, internal assistants, developer tools, and CI/CD pipelines. For each integration, organizations need to document the model used, its infrastructure, input and output data, and the responsible team.

Thales Imperva offers an AI Discovery Playbook to help organizations conduct a self-audit of their AI census, assess risks, and implement appropriate governance and protection measures.

In summary, managing AI integrations requires a structured approach, combining accurate discovery, risk assessment, and robust governance measures to ensure compliance and operational resilience.

More videos

Behind every article is the work of our passion. If you enjoy GoYou.it, you can support its development and continuity.

Support with PayPal