AI Is Breaking the Entire Structure of Software That Powers the World

Futurism ·

AI Is Breaking the Entire Structure of Software That Powers the World

The systems we take for granted are crumbling before our eyes. The post AI Is Breaking the Entire Structure of Software That Powers the World appeared first on Futurism .

AI’s ability to churn out code and find vulnerabilities is making swiss cheese out of the world’s digital infrastructure.

On Tuesday, the cybersecurity firm Calif revealed that it used a new OpenAI model to discover a critical flaw in Java, one of the most widely used programming languages in the world which undergirds a vast portion of the online and financial worlds. It’s an eerie example of how the tech is capable of breaching government and corporate systems , at a pace that may soon outstrip our ability to protect those assets — if it hasn’t already.

The bug, which the researchers are calling Java4Shell, was discovered in Java’s base64 decoder, a system that allows binary data to be represented as legible plain text. They found that Java bugs out when it tries to decode 64 “A” characters punctuated with a single equal sign. This makes it erroneously execute code, which would normally cause the program to crash by overloading the system’s memory.

But using OpenAI’s new GPT-6 cyber model, the researchers found that they could manipulate how the memory is corrupted to access the underlying system and allow a hacker to uncover sensitive data or remotely execute code.

Alarmingly, it took little effort to find. Calif says the researcher behind the discovery spent only an hour to uncover the bug, two hours to develop an exploit, and burned less than $200 in AI tokens.

Java is an all-around popular programming language, but it’s particularly prevalent among tech, corporate, and government systems. About 90 percent of Fortune 500 companies use it, the N ew York Times noted in its reporting .

Perhaps the most alarming find was that the bug could be used to break into Keycloak, a popular open source authentication system used by big organizations ranging from NASA to UnitedHealthcare.

“Java is the beating heart of the modern economy,” Bruce Dang, a senior researcher at Calif, told the NYT . “Practically speaking, all major companies” will be affected by the bug, he added.

The findings come amid a flurry of tech disruptions in recent weeks. Banks have gone down, widely used Microsoft products went offline , the nation’s flight controller system crashed, and Apple services experienced an outage . This week, hackers infiltrated the websites used by hundreds of major organizations and implanted malicious JavaScript code to steal visitors’ money and credit card details. Freezer systems used by grocery stores and blood banks have been breached by hackers in Europe . The chaos hasn’t been definitely linked to AI attacks, but it contributes to the unease around the robustness of these systems we take for granted.

Last month, Calif researchers also used an OpenAI model to find stunning security flaws in PHP , the software that underpins nearly 70 percent of all websites on the internet, in another show of AI’s threat to cybersecurity.

The Calif researchers, in a nod to the mythological creatures who are half human, half horse, say that their findings hail a new era of the “centaur hacker” — one who is part human, part AI.

Soon, though, they warn, there will come a point when an autonomous model doesn’t need a human handler to wreak havoc.

“I don’t know how much time we have — maybe three months, maybe three years,” Duong told the NYT . “But right now, the best hacker is the centaur hacker.”

More on AI: The Trump Administration Is Absolutely Flailing Trying to Explain Why So Many Crucial Services Are Going Down

The post AI Is Breaking the Entire Structure of Software That Powers the World appeared first on Futurism .

Источник: Futurism