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:
- One Step Ahead in Cyber Hide-and-Seek: Automating Malicious Infrastructure Discovery With Graph Neural Networks
January 13, 2025
When launching and persisting attacks at scale, threat actors can inadvertently leave behind traces of information. They often reuse, rotate and share portions of their infrastructure when automating their campaign’s setup before launching an attack. Defenders can leverage this behavior by pivoting on a few known indicators to uncover newer infrastructure. This article describes the benefits ...
- Analyzing CVE-2024-44243, a macOS System Integrity Protection bypass through kernel extensions
January 13, 2025
Microsoft Threat Intelligence discovered a new macOS vulnerability that could allow attackers to bypass Apple’s System Integrity Protection (SIP) in macOS by loading third party kernel extensions. SIP is a security technology that restricts the performance of operations that may compromise system integrity; thus, a SIP bypass affects the overall security of the operating system. Bypassing ...
- Deep Dive Into a Linux Rootkit Malware
January 13, 2025
This is a follow-up analysis to a previous blog about a zero day exploit where the FortiGuard Incident Response (FGIR) team examined how remote attackers exploited multiple vulnerabilities in an appliance to gain control of a customer’s system. At the end of that blog, Fortinet researchers revealed that the remote attacker had deployed a rootkit (a ...
- Nominet confirms cybersecurity incident linked to Ivanti VPN hacks
January 13, 2025
Nominet, the U.K. domain registry that maintains .co.uk domains, has experienced a cybersecurity incident that it confirmed is linked to the recent exploitation of a new Ivanti VPN vulnerability. In an email to customers, seen by TechCrunch, Nominet warned of an “ongoing security incident” under investigation. Nominet said hackers accessed its systems via “third-party VPN software ...
- UK: Hackney Council still addressing 2020 cyber attack
January 13, 2025
Hackney Council has bought a new housing management system – technology that supports local authorities manage housing – as it tries to address the damage from a cyber attack four years ago. The October 2020 cyber attack left a lasting impact on services during a housing crisis which, according to the Local Democracy Report, has seen ...
- Burning Zero Days: Suspected Nation-State Adversary Targets Ivanti CSA
January 11, 2025
Today FortiGuard Labs is releasing this blog post about a case where an advanced adversary was observed exploiting three vulnerabilities affecting the Ivanti Cloud Services Appliance (CSA). At the time of our investigation, two out of the three identified vulnerabilities were not publicly known. This incident is a prime example of how threat actors chain zero-day ...