Pular para o conteúdo
Tecnologia · Mundo

Calif uses AI to uncover critical Java flaw affecting global systems

The Java4Shell vulnerability could let attackers gain remote access to internal systems, prompting urgent patches from Oracle and government agencies.

Por Redação

Fonte: InfoMoney

4 min de leitura

Foto: Reproducao / InfoMoney
Foto: Reproducao / InfoMoney

Leitura em voz alta com a voz do seu aparelho, em português do Brasil.

AI‑driven vulnerability hunt

A cybersecurity startup revealed it used a new artificial‑intelligence model to find a critical software flaw in Java, one of the world’s most widely used programming languages. The discovery highlights a growing trend where AI accelerates the identification of security weaknesses far faster than traditional methods. According to the InfoMoney portal, the startup, Calif, leveraged its GPT‑6 Cyber model, a recent release from OpenAI, to detect the issue. The language powers roughly 90 % of Fortune 500 companies and many government agencies, including those tied to national security.

Researchers at Calif said the Java vulnerability could allow remote attackers to access sensitive internal systems. Bruce Dang, a senior researcher at Calif, described Java as “the beating heart of the modern economy” and warned that almost every large enterprise could be affected. The flaw, later named Java4Shell, was found in a component responsible for base‑64 decoding, a common algorithm for converting binary data into readable text.

The bug triggers when Java processes a sequence of 64 “A” characters followed by an equals sign. Under certain conditions, the error corrupts memory, causing the program to crash and, with the help of AI, enabling a technique called heap grooming that gives attackers full unauthorized system access. A security researcher at Calif, Quang Luong, discovered the issue using the AI model. The New York Times independently verified the attack after receiving a live demonstration from Calif, and the test succeeded in under an hour.

Technical details and impact

Calif demonstrated that the vulnerability could be weaponized against Keycloak, a popular authentication system used by thousands of organizations, including CERN, Hewlett Packard, UnitedHealthcare and NASA. The demonstration showed how an attacker could exploit the flaw to break into internal networks, a scenario that could have far‑reaching consequences for critical infrastructure. Thai Duong, CEO of Calif, noted that while the flaw is not trivial to turn into a practical attack, it still poses a long‑term threat.

He warned that unpatched systems could remain exposed for five to ten years, a period during which attackers could refine exploitation methods. The vulnerability’s root lies in Java’s handling of base‑64 strings, a routine operation in many software stacks. The error leads to memory corruption that typically crashes the application, but AI‑assisted heap grooming can manipulate memory to achieve full system compromise.

Responses from industry and government

Oracle, the owner of the Java brand, issued a security update on the same day the flaw was disclosed. The company did not provide additional commentary, but the patch is intended to correct the base‑64 decoding issue. The U.S. National Security Agency (NSA) described a network outage affecting its classified systems as routine maintenance, and it declined to comment on any possible link to the newly discovered vulnerability.

The disclosure follows a pattern of AI‑aided findings by Calif, which also identified serious security holes in PHP the previous month. Those PHP flaws prompted governments, including the NSA, to advise companies to take down websites if immediate patching was not possible. Although the eventual impact was milder than expected, the PHP episode underscored how small research teams can now uncover issues that previously required large, state‑level resources.

Ollie Whitehouse, chief technology officer of the United Kingdom’s National Cyber Security Centre, said the trend of AI‑driven vulnerability research is reshaping the security landscape. He noted that the same technology that helps defenders also lowers the barrier for malicious actors. The rise of “centaur hackers”—a blend of human expertise and AI power—means that sophisticated attacks can be crafted in days or hours rather than months of elite effort.

Future threats and policy concerns

Former world chess champion Garry Kasparov’s concept of a “centaur” player—human plus machine—now applies to cybersecurity. Experts warn that the best hackers may soon be AI‑augmented teams capable of launching complex campaigns at scale. Thai Duong believes it is only a matter of time before AI can autonomously develop the most advanced attacks, ushering in a new era of digital risk.

The FBI issued a warning this week that hackers are already using autonomous AI tools to target government agencies and corporations worldwide. Duong cautioned that the window for effective defense may be measured in months or years, not decades. “At this moment, the best hacker is the hacker centaur,” he said, emphasizing the need for rapid patching and continuous monitoring.

The incident also reignited debate over responsible disclosure and the role of AI in security research. While AI can surface flaws quickly, it also raises concerns about unintended consequences if the same tools fall into the wrong hands. Industry observers stress that coordinated response—prompt patches, clear communication, and shared threat intelligence—is essential to mitigate the fallout from AI‑discovered vulnerabilities.

Leia também

Comentários

    Deixe seu comentário

    Seja respeitoso. Não aceitamos ofensas, links ou propaganda. Veja os Termos de uso.