A bug in Java, one of the most widely used programming languages in the world, can be exploited to take over computer systems, and researchers say it was found with help from artificial intelligence.

Researchers at Calif, a cybersecurity startup, discovered the flaw with the help of OpenAI’s GPT-6 Cyber model, which is not yet publicly available. The bug is triggered in a surprisingly simple way. On some machines, feeding Java’s base64 decoder a string of 64 “A” characters followed by an equals sign causes a memory error. Usually, that error just crashes the program. But when the flaw is manipulated in specific ways, it can give an attacker full unauthorized access to the system.

To show how serious the problem could be, Calif used the bug to break into Keycloak, a widely used login system that protects organizations such as CERN and NASA. Java is embedded across the economy. It powers platforms for banks, hospitals, airlines and government agencies, and it is used by about 90 percent of Fortune 500 companies.

Calif says the bug is not easy to weaponize, which should give defenders some time. Still, the timing leaves little room for complacency. Oracle shipped a critical patch update on Tuesday, but it did not specifically mention this bug. The Cybersecurity and Infrastructure Security Agency, or CISA, has told organizations to patch their systems immediately.

Patching alone may not close the gap quickly. Some systems will inevitably go unpatched, and Calif estimates the bug could remain in use for five to 10 years.

The disclosure comes after a rough week for digital infrastructure. Outages were reported at Microsoft, the Federal Aviation Administration, banks, Apple and a range of other services. Officials have not publicly linked those disruptions to AI, and the Java flaw is a separate issue. Still, the events point to aging digital systems that may not be ready for the threats now emerging.

The episode also highlights a shift in how security flaws are found. A string of A’s that looks like nothing more than a typo is, in this case, a way into unpatched computers. As AI tools are used to probe software for weaknesses, security teams may find that the same tools that help them defend systems can also expose the gaps attackers are looking for.

Este artículo fue escrito con la asistencia de IA.
News Factory APP - noticias agénticas para impulsar tu SEO y AEO.