Java4Shell Vulnerability: AI-Found OpenJDK Flaw Gets a Patch

Summary
Security start-up Calif says it used AI to find Java4Shell, a critical memory-corruption flaw in OpenJDK’s Base64 decoder. Oracle’s November 17 security update patches it. California rates the bug 9.2 out of 10, and the firm says Java runs at 90% of Fortune 500 companies.
Key Highlights
- Tracked as CVE-2026-114938, the Java4Shell vulnerability lets malformed input overwrite neighbouring memory.
- Calif says one researcher found it in about an hour and built an exploit in two more.
- The firm puts the AI token cost under $200.
- It says exploitation depends on specific hardware.
What Happened?
California disclosed the OpenJDK flaw after Oracle shipped its fix. Careful exploitation, the firm says, can expose sensitive memory or allow remote code execution. California earlier disclosed PHP2Shell, a flaw it also found with AI help.
Why Is It Important?
Java’s memory safety is one of its main selling points, so a memory-corruption bug in a core component is unusual. Banks, hospitals, airlines, and government agencies rely on the platform. The Java4Shell vulnerability also shows how cheaply AI tools can now speed up bug-hunting and exploit-building.
Impact and Implications
Defenders face a compressed window between disclosure and attacks. GreyNoise says it is monitoring scanning and exploitation attempts against the flaw. Severity scores also differ by source: Calif says 9.2, while German outlet heise reports 8.1.
What Happens Next?
Organisations running OpenJDK should apply Oracle’s November update and check vendor advisories for bundled Java runtimes. Watch for further exploitation reports and any clarification on the final severity rating of the Java4Shell vulnerability.

