- 99% of organizations have sensitive data exposed to AI tools.
- 90% of sensitive cloud data remains accessible to AI tools.
- 98% of organizations have unverified apps, including shadow AI.
A new report from Varonis (VRNS, Financial) has brought to light significant security vulnerabilities associated with AI integrations across 1,000 organizations. The study, which covered nearly 10 billion cloud resources and over 20 petabytes of data, reveals that sensitive data is exposed to AI tools in 99% of the cases analyzed.
The data risk assessment further indicates that 90% of sensitive cloud data, crucial for AI training, is unnecessarily accessible to AI tools, underscoring severe security gaps within most organizational ecosystems. Additionally, the report divulges that 98% of these organizations' systems include unverified apps, many of which are categorized as shadow AI, while 88% have ghost users active within their environments.
Interestingly, 1 in 7 organizations fails to enforce Multi-Factor Authentication (MFA) across their cloud environments, a basic yet critical security measure. These findings emphasize the urgent need for a data-centric approach to security, especially as organizations accelerate AI adoption under pressures highlighted by Varonis CEO Yaki Faitelson.
Varonis's report establishes the company as a leader in the field of AI security, presenting its platform as a vital solution for mitigating these risks. The findings aim to raise awareness and prompt actionable responses from organizations to protect sensitive data against emerging AI-related threats.
