Attackers can indirectly instruct AI for malicious aims
We show you that LLM-integrated apps can suffer indirect prompt injection attacks through different methods that can put the security of their users at risk.
Let's rather say a bunch of breaches in a single box
They said they discovered the mother of all data breaches. But it was a false alarm. Let's understand why.
NIST sheds light on the classification of attacks on AI
Here is an overview of a recent NIST report on adversarial machine learning that could help us understand more about attacks against and from AI systems.
Outstanding incidents of this type of crime since 2011
Discover some of the most notorious crypto crimes to date, among which we present not only software cyberattacks but also tremendous scams.
Introduction to a trend of just over a decade ago
We share basic knowledge about cryptocurrency and related concepts, including crypto crime, a worrying issue that has gone hand in hand with that paradigm.
Among exposed were secrets, code and AI training data
We describe the leak and mention its causes and threats, talk about the possible security risks when using Azure SAS tokens and give advice to prevent leaks.
A brief overview of this recent EU draft regulation
Learn about the draft Cyber Resilience Act, including its purposes, obligations and penalties, before its final version comes.
Increase the board's cyber savvy with these reads
Boards need to gain skills to identify, assess and supervise their organizations' cyber risk-management measures. We share resources to help them on their way.
Soon it will be a must in cybersecurity due to NIS2
We briefly introduce the NIS 2 Directive and pay special attention to its cybersecurity training requirement for organizations' boards of directors.