AI & software security
-
Oracle ships 673 enterprise security patches
Oracle’s September Critical Security Patch Update contains 673 new patches spanning databases, middleware, business applications, analytics, Java, and other enterprise technology.
-
Compromised plugin vendor backdoors WordPress sites
Attackers compromised the distribution infrastructure for Admin Menu Editor Pro and pushed backdoored updates to around 230 customers, affecting at least 1,500 WordPress sites.
-
Hackuity raises $19m for vulnerability operations
French cybersecurity company Hackuity has raised $19 million to expand its vulnerability-operations platform as enterprises contend with growing volumes of security findings.
-
Spain records first AI-agent breach notification
Spain’s data regulator has received its first breach notification involving an AI agent, with the reported attack chaining authenticated access, vulnerability discovery, and access to personal information.
-
Microsoft sets cyber limits for future AI
Microsoft AI has published a draft code intended to govern future MAI models, including explicit limits on offensive cyber activity, autonomy, privilege, and resistance to human shutdown.
-
Exein raises $270m for physical AI security
Rome-based Exein has raised $270 million at a $1.7 billion valuation as it expands embedded security technology for robots, vehicles, industrial systems, and other AI-enabled machines.
-
ENISA details CRA platform security testing
ENISA has confirmed that AI-based secure-code review formed part of security testing for the EU platform now receiving mandatory Cyber Resilience Act vulnerability and incident reports.
-
Lagarde warns of Europe’s AI dependence
ECB President Christine Lagarde has warned that Europe’s reliance on imported AI technology could become a strategic vulnerability as the technology spreads through essential services.
-
Europe weighs controls for autonomous AI
European governments are weighing new safeguards for frontier AI as systems capable of independently reaching external infrastructure test the limits of existing oversight.
-
Artifactory attacks move into persistent compromise
New telemetry shows attackers chaining JFrog Artifactory vulnerabilities to obtain administrative control, establish persistence, execute code, and deploy backdoors across compromised repository servers.