Summary
This report analyzes data security risks associated with AI across various organizations. It details alarming findings from a review of 1,000 IT environments, revealing that 90% of organizations have exposed sensitive cloud data. The report also highlights issues related to ghost users, with 88% having stale but enabled accounts that can pose security threats. Furthermore, 98% of organizations utilize unverified applications, emphasizing the risks posed by shadow AI. The report discusses the significant challenge of managing permissions for sensitive data in the cloud, particularly concerning AI training data. Notably, model poisoning risks are outlined, demonstrating how attackers can corrupt AI models. The analysis includes data from diverse industries and geographic regions, underscoring the universal nature of these security challenges. The report emphasizes the need for robust identity management practices and vigilant oversight to mitigate the risks of data exposure and breaches in AI-driven environments.
