_large

AI is now finding bugs in the code that runs banks, airlines, and hospitals

An AI model created for security research has uncovered a chain of weaknesses in Java, the Oracle language used across airlines, banks, hospitals, and government systems. The flaws were identified by Calif using OpenAI’s GPT-6 Cyber model, prompting CISA to call on companies to update every system running Java. Oracle issued a critical security patch update the same day and urged customers to install it, though it did not elaborate further. For many organizations, the fix may be difficult to carry out, given that Java is embedded deep within the software used by major enterprises, with 90% of Fortune 500 companies relying on it.

Bruce Dang of Calif said Java’s reach makes the issue especially serious, describing it as essential infrastructure and noting that large companies will have significant work ahead to secure their systems. Calif said it alerted Oracle last month and then coordinated with government agencies through the Five Eyes alliance, while Amazon, Apple, Google, and Microsoft also assisted in arranging fixes. The aim of that coordination is to have patches ready before technical details spread widely enough for attackers to take advantage.

The Java discovery follows a series of similar findings enabled by AI in recent months. In October, researchers using OpenAI and Google models identified two PHP vulnerabilities that had gone unnoticed for more than a decade, leading companies and government bodies to take websites offline as a precaution. The cases reflect a broader change in security research, where AI can scan large and mature codebases and detect flaws that human review missed. But the same tools that help defenders could also help attackers, intensifying the race to find and repair weaknesses first. The current outages affecting digital services in the US have not been linked to Java, yet the timing adds pressure on organizations to locate every instance of the language and apply Oracle’s fix.

Read More