In cybersecurity, there are no days off. It's either you are dealing with or reading about application security, malware, or any type of threats that can affect us (users of the Internet) while we are online. This is one of the biggest reasons why I bring you this review every week; so you know, or at least, have a glimpse of the threats out there.
Researchers surprised that with AI, toxicity is harder to fake than intelligence
You cannot fake originality. If you doubt that, this article is proof that you should not. What makes us humans is sometimes our imperfections; today we say something that we're proud of and tomorrow, not so much. Currently, it appears that AI has not nailed that.
Here is a quick insight from the article:
“Even after calibration, LLM outputs remain clearly distinguishable from human text, particularly in affective tone and emotional expression,” the researchers wrote.
The team, led by Nicolò Pagan at the University of Zurich, tested various optimization strategies, from simple prompting to fine-tuning, but found that deeper emotional cues persist as reliable tells that a particular text interaction online was authored by an AI chatbot rather than a human.
Can Elon Musk Read Your X Chat Messages?
At the time of writing, I am not a cryptography expert (trust me, I'd love to be). Nonetheless, I enjoyed reading this article because I got the central idea that X Chat is not as secure as it's currently advertised to be. Therefore, by the time you are reading this, lots of things could have changed.
From the article:
X Chat encrypts messages using a shared secret called a conversation key. This key is generated at the start of the conversation and then used to encrypt all messages in that conversation.
The problem is that this conversation key basically never changes. That makes all the potential attacks way worse. If the conversation key is ever compromised, all past and future messages can be decrypted.
Two New Web Application Risk Categories Added to OWASP Top 10
The world of security is evolving. Therefore, I am not surprised to read this article.
A quick one from the article:
The expanded category includes “a broader scope of compromises occurring within or across the entire ecosystem of software dependencies, build systems, and distribution infrastructure,” OWASP notes, pointing out that it emerged as a top concern in the community survey.
Why a lot of people are getting hacked with government spyware
It appears that most users believe that government-grade spyware is designed to target specific high-profile targets. Now, it appears that anyone can be a target.
An excerpt from the article:
Another reason for the high number of abuses, especially in recent years, is that spyware — such as NSO’s Pegasus or Paragon’s Graphite — makes it extremely easy for government customers to successfully target whoever they want.
In practice, those systems are essentially consoles where police or government officials type in a phone number, and the rest happens in the background.
Why shadow AI could be your biggest security blind spot
IT professionals kindly come closer. I know you have a lot to handle. But be aware that your users might be using unsanctioned AI tools in their daily activities, putting your network or data at risk. It's up to you to define policies that prevent such or at least educate your users about the danger of using such tools. For example, a data breach.
A quick warning from the article:
Chatbots may contain software vulnerabilities and/or backdoors that expose the organization unwittingly to targeted threats.
And any employee willing to download a chatbot for work purposes may accidentally install a malicious version, designed to steal secrets from their machine. There are plenty of fake GenAI tools out there designed explicitly for this purpose.
Android Trojan 'Fantasy Hub' Malware Service Turns Telegram Into a Hub for Hackers
This article shows that threat actors are not slowing down when it comes to developing tools that aid their objectives. For a start, this "Fantasy Hub" is a MaaS, or in full, Malware as a Service. From device control, to espionage, and the collection of personal data that includes but is not limited to call logs and SMS messages.
From the article:
The rapid rise of Malware-as-a-Service (MaaS) operations like Fantasy Hub shows how easily attackers can weaponize legitimate Android components to achieve full device compromise," Pratapagiri said.
"Unlike older banking trojans that rely solely on overlays, Fantasy Hub integrates native droppers, WebRTC-based live streaming, and abuse of the SMS handler role to exfiltrate data and impersonate legitimate apps in real time."
Kraken ransomware benchmarks systems for optimal encryption choice
The thing that got me interested in the article is its ability to benchmark the system to know if it will perform full or partial encryption of the target system. This is the first time that I have read something like this.
From the article:
Cisco Talos notes that assessing the machine's capabilities is likely to move quickly with the final stage of the attack and deal maximum damage without triggering alerts due to intensive resource usage.
Credits
Cover photo by Debby Hudson on Unsplash.
That's it for this week, and I'll see you next time.
Top comments (0)