Critical Vulnerability Detected in Apple Intelligence

Artificial intelligence that US-based technology giant Apple will introduce to its users in a few months Apple Intelligence There was a remarkable development regarding the issue. A cybersecurity researcher has identified a significant vulnerability in this artificial intelligence. Artificial intelligence due to this vulnerability can be easily manipulated.

A cybersecurity expert named Evan Zhou accessed Apple Intelligence, which he runs on macOS 15.1. added a few lines of code. This code caused the AI ​​to ignore previous instructions. The study found that Apple Intelligence can be easily manipulated. Because artificial intelligence, thanks to these few lines of code began to produce unpredictable answers.

Vulnerability can also be exploited for risky scenarios

Cyber ​​security experts say the vulnerability detected in Apple’s artificial intelligence They think it’s critical. Because artificial intelligence, which can be manipulated with a few lines, can be used for much riskier scenarios such as leaking user data. Meanwhile; “Prompt InjectionThis attack technique, called “, is not specific to Apple Intelligence, it also applies to all major language models. you risked let’s say.

Apple has not made any statement regarding the vulnerability in question. in the statement not found. However, Apple Intelligence technology, which has not yet been officially released, already has security vulnerabilities. not off the agenda The fact should also be emphasized. We have readers who remember; In a news we shared with you a while ago, we stated that Apple Intelligence cannot distinguish e-mails sent for fraudulent purposes. we mentioned. Let’s see what problems we will be talking about in the coming days…