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:
- Portland-based BerryDunn faces lawsuits over data breach affecting 1.1M people
May 6, 2024
A Maine accounting firm has been sued after a data breach exposed the personal information of more than a million people. BerryDunn, an accounting firm based out of Portland, is facing lawsuits from nine different people. The lawsuits claim that names and Social Security numbers of 1.1 million people were exposed in the breach involving one ...
- Germany recalls envoy to Russia over cyberattack
May 6, 2024
The German ambassador to Russia was recalled for consultations on Monday after Berlin accused Moscow of carrying out cyberattacks. A newly concluded government investigation found the cyberattack had been carried out by a group — linked to Moscow’s GRU military intelligence agency — known as APT28. The group, also known as Fancy Bear, has been accused ...
- Singapore law firm Shook Lin & Bok hit by cyber attack; allegedly paid ransom
May 5, 2024
Singapore law firm Shook Lin & Bok was hit by a ransomware attack in April, and the incident is now under investigation by the local authorities. In response to queries from The Straits Times, the firm said in a statement on May 2 that the incident was discovered on April 9, and it immediately engaged a ...
- U.K., U.S. and Canadian cyber authorities warn of pro-Russia hacktivist attacks on OT systems
May 3, 2024
The U.K.’s National Cyber Security Centre (NCSC) and other international cyber authorities, including the Federal Bureau of Investigation (FBI), have warned about pro-Russia hacktivist attacks targeting providers of operational technology. OT is hardware and software that interacts with the physical environment and includes smart water metres, automated irrigation systems, dam monitoring systems, smart grids and IoT ...
- Graph: Growing number of threats leveraging Microsoft API
May 2, 2024
An increasing number of threats have begun to leverage the Microsoft Graph API, usually to facilitate communications with command-and-control (C&C) infrastructure hosted on Microsoft cloud services. The technique was most recently used in an attack against an organization in Ukraine, where a previously undocumented piece of malware used the Graph API to leverage Microsoft OneDrive for ...
- Watch out for tech support scams lurking in sponsored search results
May 2, 2024
A campaign using sponsored search results is targeting home users and taking them to tech support scams. Sponsored search results are the ones that are listed at the top of search results and are labelled “Sponsored”. They’re often ads that are taken out by brands who want to get people to click through to their website. ...