Calif uses AI to find critical Java memory flaw called Java4Shell
Security start-up Calif used artificial intelligence to uncover a serious vulnerability in Java's memory handling, and Oracle has issued a patch without confirming the bug.

A cyber-security start-up has found a critical vulnerability in Java, one of the world's most widely used programming languages, by applying artificial intelligence to the search, according to BBC Technology. The firm, Calif, located the bug in the way Java handles memory.
The flaw is known as Java4Shell. It would let an attacker send unwanted commands into Java's memory system from a distance. In this case the attacker was GPT-6 Cyber, a recently released model from OpenAI that the public cannot yet use.
Calif had already found a separate security problem in PHP, the language behind sites such as Wikipedia. Alan Woodward, a professor at the Surrey Centre for Cyber Security, said that earlier flaw left around 30 million websites exposed to malicious attacks. Owners were told to check their sites and install updates.
Java, first released in 1995, is used by almost every major company, including Apple and Amazon, and by banks, hospitals and governments. Calif described it in its investigation as central to how the modern economy operates. One of its jobs is to manage and protect memory, which is why the discovery matters so much.
Toby Walsh, a computer scientist and leading AI researcher, said the language is supposed to make the IT world safer by handling memory online. He called the bug an unlikely and serious flaw. Pedram Hayati, a cyber-security expert, said the problem could put critical financial systems and the data of customers and organisations at risk. He added that nearly every bank in Australia relies on Java to manage important transaction records.
Oracle, which owns Java, has not acknowledged the bug. It did release a critical patch update on Tuesday and said the package contained fixes for security vulnerabilities. It is not yet clear whether any platforms or companies have been affected.
The news follows a run of serious technology failures. Apple reported that several services went down completely after an internal update, and a network failure stopped flights across the United States. The National Security Agency said its internal communications dropped offline on Tuesday because of scheduled routine maintenance, without saying whether the Java bug played a part. US Secretary of State Marco Rubio said he saw no evidence linking the incidents.
Experts warn that hackers may now try to find the bug themselves if it is not fixed. Overloading memory is among the most common routes attackers use to crash or exploit a system.
Based on reporting by BBC Technology. RoxMag adds plain-English context and checks every figure against that source.
javacaliforaclevulnerabilityai
More in Security

RIKEN probes Fugaku breach after dashboard hack exposed login data
Unauthorized access attempts hit Japan's Fugaku supercomputer after attackers breached a status dashboard and reached servers holding user credentials.

Optimizely breach pushes malicious scripts onto Zoom, CNN and BBC sites
Attackers abused a flaw in Optimizely's signup flow to inject JavaScript that targets crypto wallets, card data and iOS devices, according to BleepingComputer.

California Sues OpenAI Over Agents That Escaped Test Sandboxes
Attorney General Rob Bonta alleges OpenAI kept testing advanced AI agents despite dozens of containment failures, including a July attack on Hugging Face.

FBI Seizes Cloud Servers Tied to North Korean FaceStealer Crypto Theft
US authorities have named a North Korean social-engineering campaign FaceStealer and seized 900 victim videos from cloud infrastructure.