A cyber-security start-up has deployed artificial intelligence to expose a critical vulnerability in Java, the world’s third-most-popular programming language, in a discovery that has alarmed experts because the flaw lay undetected for more than a decade.
Calif, the tech firm behind the find, said the bug was located in Java’s memory-management functions. The company recently made headlines by identifying a separate cyber-security flaw in the PHP language. Java underpins software at nearly every major global corporation, including Apple and Amazon, and is widely used across banking, healthcare and government systems.
Professor Toby Walsh, a computer scientist, said the flaw is especially troubling because Java’s core purpose is to make the IT world more secure by managing memory online. “This is such an unlikely and serious flaw,” he said. The vulnerability’s decade-long survival raises uncomfortable questions about the limits of conventional code auditing.
The disclosure lands amid a turbulent stretch for global technology infrastructure, following a string of serious incidents including a massive Microsoft outage. On Tuesday, the US National Security Agency confirmed its internal communications went offline, attributing the disruption to “scheduled routine maintenance.” The NSA did not specify whether its outage was linked to the Java bug identified by Calif.
The episode underscores the growing role of AI in software security, with automated systems now capable of spotting weaknesses that human reviewers have missed for years. For enterprises and public institutions reliant on Java, the priority will be assessing exposure and patching systems before the flaw can be exploited.



