Plainly speaking, Artificial intelligence is no longer a fringe technology. It has become a core component of modern business, from customer service chatbots to complex data analysis.
We often treat the Large Language Models (LLMs) that are at the core of this technology as trusted black boxes. But like any software, they can be tampered with, manipulated, and turned against their creators. Understanding the ways an AI model can be compromised is the first step toward building a secure and resilient AI infrastructure. There are three primary ways that the authors want to cover, in which they will explain how an AI model can be compromised and made to act in unpredictable (or malicious!) ways:
Read more…
Source: Trend Micro
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- Proof-of-Concept Exploit Released for CVE-2024-53691 in QNAP QTS and QuTS NAS
January 17, 2025
QNAP has released a security advisory addressing three vulnerabilities in the QTS and QuTS products. QTS and QuTS are the operating system for QNAP Network-attached storage (NAS) appliances. CVE-2023-39298 is a ‘Missing authorisation’ vulnerability with a CVSSv3 score of 7.8. If exploited, a local attacker with low privileges could access data or perform actions without proper ...
- CIA employee pleads guilty over leak of classified Israeli plans
January 17, 2025
A CIA employee who was accused of leaking classified documents about Israel’s plans to strike Iran pleaded guilty on Friday to criminal charges that he willfully retained and transmitted national defense information, the U.S. Department of Justice said. In pleading guilty, Asif William Rahman, who worked at the U.S. intelligence agency since 2016, acknowledged that he ...
- Mercedes-Benz Head Unit security research report
January 17, 2025
This report covers the research of the Mercedes-Benz Head Unit, which was made by Kaspersky team. Mercedes-Benz’s latest Head Unit (infotainment system) is called Mercedes-Benz User Experience (MBUX). The researchers performed analysis of the first generation MBUX. MBUX was previously analysed by KeenLab. Their report is a good starting point for diving deep into the MBUX ...
- Threat Brief: CVE-2025-0282 and CVE-2025-0283
January 16, 2025
On Jan. 8, 2025, Ivanti released a security advisory for two vulnerabilities (CVE-2025-0282 and CVE-2025-0283) in its Connect Secure, Policy Secure and ZTA gateway products. This threat brief provides attack details that we observed in a recent incident response engagement to provide actionable intelligence to the community. These details can be used to further detect current ...
- Clop ransomware gang names dozens of victims hit by Cleo mass-hack, but several firms dispute breaches
January 16, 2025
The prolific Clop ransomware gang has named dozens of corporate victims it claims to have hacked in recent weeks after exploiting a vulnerability in several popular enterprise file transfer products developed by U.S. software company Cleo. In a post on its dark web leak site, seen by TechCrunch, the Russia-linked Clop gang listed 59 organizations it ...
- New Star Blizzard spear-phishing campaign targets WhatsApp accounts
January 16, 2025
Star Blizzard’s new spear-phishing campaign, while novel in that it uses and targets WhatsApp for the first time, exhibits familiar spear-phishing TTPs for Star Blizzard, with the threat actor initiating email contact with their targets, to engage them, before sending them a second message containing a malicious link. The sender address used by the threat actor ...