Skip to content

CISA Urges Java Updates After AI Uncovers Flaws in Critical Systems Worldwide

CISA Urges Java Updates After AI Uncovers Flaws in Critical Systems Worldwide

Researchers at cybersecurity startup Calif have found a series of vulnerabilities in Java, the Oracle-owned language that runs systems at airlines, banks, hospitals, and government agencies. The team used OpenAI’s GPT-6 Cyber model to find the flaws. The Cybersecurity and Infrastructure Security Agency is telling organizations to update every system that runs Java.

Oracle released a patch it called a “critical security patch update” on the same day and urged customers to apply it. The company did not comment further. Installing the fix may take considerable effort, since 90% of Fortune 500 companies use Java and it is deeply embedded in the software large organizations rely on.

Calif senior researcher Bruce Dang said Java’s footprint is what raises the stakes. “Java is basically critical infrastructure,” he told The Financial Times. With so much of the modern stack built on it, he said, remediation will be a heavy lift for many large enterprises.

Calif disclosed the flaw to Oracle last month, then coordinated with government agencies through the Five Eyes intelligence-sharing group (Australia, Canada, New Zealand, the UK, and the US). Amazon, Apple, Google, and Microsoft helped line up fixes as well. It’s a standard coordinated-disclosure play: get patches ready before the details circulate widely enough for attackers to weaponize them.

Java isn’t the first time this has happened lately. AI models keep turning up software flaws, and back in October, researchers used models from OpenAI and Google to spot two security holes in PHP, which runs about 70% of websites. Those bugs had been sitting there for over ten years without anyone noticing. Once they were made public, a number of companies and government agencies took sites offline just to be safe.

Together, these cases mark a shift in vulnerability research. Models can now scan large, mature codebases and surface bugs that survived years of manual review. Defenders get a head start, but attackers with the same tooling are close behind, so the window to patch is shrinking.

On one side, researchers hunt for weaknesses so they can patch them in time. On the other, malicious actors hunt for the same weaknesses to exploit them. AI has sharpened the contest, because the tools that help one side can help the other.

The warning from CISA came at an uneasy time. Digital services across the US had been faltering for weeks. Last week, Microsoft went down, and the company traced it to a system that failed during a routine effort to reroute user traffic. No cyberattack, it said. JPMorgan and United Airlines had stumbled too, and the worry was that more outages were coming.

Aviation took a hit too. The aging US air traffic control system went down for five hours, snarling travel for a full weekend. Elsewhere, Norwegian energy group Equinor disclosed that it had been hit by a cyberattack, though it’s unrelated to the air traffic outage.

There’s no evidence tying the outages to the Java bugs, but the timing raises the stakes for Java-heavy shops. With Oracle’s patch out and CISA’s warning public, the job now is inventory and remediation: find every system with Java embedded in it, then patch each one.

Maybe you would like other interesting articles?

Leave a Reply

Your email address will not be published. Required fields are marked *