Protect AI reports that 32 security defects were discovered in multiple open source artificial intelligence and machine learning tools over the past few months. These defects were reported via Protect AI’s Huntr AI bug bounty platform. Some of the bugs could cause high levels of harm, such as information loss, access to restricted materials, and server takeover. One of the vulnerabilities, CVE-2024-22476, could allow remote attackers the access and ability to escalate privileges. Other flaws could be exploited to cause denial-of-service (DoS) conditions, the overwriting of files, and potentially a full takeover. Protect AI is continuing to work with maintainers to fix the reported vulnerabilities.