A recent wave of cybersecurity incidents has raised concerns about the vulnerability of sensitive data and the growing threat of AI-driven attacks. In one notable incident, a critical vulnerability chain dubbed SearchLeak in Microsoft 365 Copilot Enterprise was discovered, allowing attackers to steal sensitive data from a target's mailbox, OneDrive, or SharePoint account through a specially crafted URL. This vulnerability was addressed by Microsoft at the beginning of the month, but not before it was exploited by attackers.

Meanwhile, the open-source AI orchestration platform Langflow has been targeted by attackers exploiting a high-severity path traversal flaw, despite a patch being available for over two months. This vulnerability can allow attackers to write files to arbitrary locations within the affected system and, under certain conditions, can be used to achieve remote code execution (RCE) on affected servers.

In a separate incident, the ShinyHunters extortion gang stole personal information from over 137,000 school staff accounts in a Salesforce data theft attack that targeted the widely used Infinite Campus K-12 student information system in March.

These incidents highlight the growing concern of data breaches and AI-driven threats, which can have serious consequences for individuals and organizations. The use of AI models, such as those employed by Microsoft 365 and Langflow, can amplify the impact of these attacks, allowing attackers to quickly and easily access sensitive information.

"The use of AI models has made it easier for attackers to automate their attacks and evade detection," said Jim Sherlock, VP of cybersecurity R&D at ProCircular. "This has created a new level of risk for organizations, which must now contend with both traditional and AI-driven threats."

• **137,000: The number of school staff accounts affected by the Infinite Campus data breach
• **2 months: The length of time a patch has been available for the Langflow vulnerability
• **11 million: The number of students whose data is managed by Infinite Campus
• **46 states: The number of states in which Infinite Campus operates

Experts warn that the convergence of AI models and data breaches creates a perfect storm of risk for organizations. "The use of AI models has created a new level of risk for organizations, which must now contend with both traditional and AI-driven threats," said Sherlock.

"The key to mitigating this risk is to implement robust security measures, including behavioral AI, to detect and prevent these types of attacks," said Dan Nickolaisen, Solutions Architect Manager at Abnormal AI.

The use of AI models has become increasingly prevalent in recent years, with many organizations employing these models to automate tasks and improve efficiency. However, this increased use of AI has also created new risks, as attackers have begun to exploit vulnerabilities in these models to gain access to sensitive information.

As the threat of AI-driven attacks continues to grow, organizations must take steps to mitigate this risk. This includes implementing robust security measures, such as behavioral AI, to detect and prevent these types of attacks. Additionally, organizations must prioritize the security of their data, ensuring that sensitive information is properly protected and that vulnerabilities are quickly addressed.

• Who: Microsoft, Langflow, Infinite Campus
• What: Data breaches and AI-driven attacks
• When: Recent incidents have occurred in the past few months
• Impact: Sensitive information has been compromised, and the risk of AI-driven attacks has been highlighted