Article
AI Deep Dive Part 3: Understanding Biases & How Threat Actors Use AI
Arete Analysis
Cybersecurity 101
In the final installment of Arete’s Artificial Intelligence (AI) deep dive, we will explore different types of biases that can be present in AI and how threat actors are able to leverage AI in cybercrime. Inherent biases are present in both human thought processes and AI models, which often influence the data used for training and the algorithms themselves. Additionally, while many AI tools are designed and intended for legitimate purposes, users should be aware that they are also continuously used by cybercriminals to enable operations.
Biases in AI
Humans all have inherent biases in our thought processes, which is also true in AI models. The data used to train models, and the models themselves are victims of biases which influence the final AI product and its responses. When considering biases within AI, algorithmic biases are top of mind. Algorithmic bias outlines the systematic and repeatable errors in a computer system that create unfair outcomes. A study conducted by the European Union’s Artificial Intelligence Act (EU AIA) did a great job of outlining the real-world implications of biases infiltrating AI.
Source: 5 Things You Must Know Now About the Coming EU AI Regulation
What begins with minimal risk at the bottom of this graphic, with biases influencing video gameplay and spam filters, quickly progresses to high-impact areas like transportation systems, justice systems, and even facial recognition. A real-world example of this would be the situation where an individual is wrongly accused of a crime due to biases in AI based on their age, gender, or skin color, evidence that algorithmic biases need to be understood and mitigated to correct incorrect outcomes facilitated by potential false information generated due to algorithmic bias.
AI in the Wrong Hands
Just as AI enables the work of professionals across many industries, cybercriminals have also begun to exploit this technology. While most public AI models have filters in place to prevent the malicious use of the models, these filters can often be bypassed to create convincing malicious social engineering content. Additionally, some threat actors, like the ransomware group Funksec, have developed their own AI models without these limitations.
During an interview, an excerpt of which is shown below, the leader of Funksec emphasized their ability to act as developers and bring high-level ideas while AI acts as the programmer, enabling their ideas using their proprietary WormGPT module. This AI use case means less technical cybercriminals are becoming increasingly able to write malware without the scripting knowledge typically required.
An excerpt from an interview with a Funksec operator. Source: Threat Actor Interview: Spotlighting on Funksec Ransomware Group
Here’s an example of a WormGPT response when asked for a webshell and C++ code.
Source: Threat Actor Interview: Spotlighting on Funksec Ransomware Group
Conclusion
In the right hands, AI is a powerful tool for productivity and advancement, but the potential risks associated with the growth of AI in connection with illegal activity such as cybercrime should be carefully monitored and addressed. Both the unintended negative effects of biases and the use of AI by cybercriminals highlight the challenges for evolving safeguards and controls protecting correct and ethical AI use.
Sources
5 Things You Must Know Now About the Coming EU AI Regulation
FunkSec: An AI-Centric and Affiliate-Powered Ransomware Group
Image Sources:
Back to Blog Posts
Article
Arete's 2026 Q1 Crimeware Report
Harness Arete’s unique data and expertise on extortion and ransomware to inform your response to the evolving threat landscape.
Article
CMS Vulnerability Leads to ClickFix Campaign
Threat actors compromised at least 700 education and technology websites in a recent ClickFix campaign by exploiting a critical SQL injection flaw (CVE-2026-26980) in the Ghost content management system (CMS). Adversaries combined the vulnerability with the ClickFix social engineering tactic to steal admin keys and inject a malicious JavaScript that delivers a fake Cloudflare or CAPTCHA verification pop-up, tricking victims into copying and pasting a malicious command into their systems.
What’s Notable and Unique
Rather than targeting the end user first, this campaign is unique in its initial exploitation of the system, followed by social engineering attempts. This hybrid attack style is likely being leveraged to bypass traditional defenses.
This recent campaign also highlights how trusted web properties can be weaponized at scale and coupled with unpatched CMS vulnerabilities. Rather than using the CMS compromise to perpetrate a single attack, threat actors turned it into a supply-chain attack that ultimately affected over 700 trusted websites.
Analyst Comments
As network defenders and their tools enhance threat detection capabilities, adversaries increasingly seek methods to bypass these defenses. By combining vulnerability exploitation, social engineering techniques, and staging for ancillary attacks, this campaign successfully bypassed traditional defenses and inflicted significant impact. Defending against hybrid cyberattacks requires comprehensive security controls beyond simply patching vulnerabilities. Organizations should focus on limiting movement within the environment, detecting abuse of trusted applications, and preventing end-user manipulation.
Sources
700+ education and tech websites hijacked in huge ClickFix malware campaign
Under the engineering hood: Why Malwarebytes chose WordPress as its CMS
Think before you Click(Fix): Analyzing the ClickFix social engineering technique
Ghost CMS Vulnerability Exploited to Infect 700 Sites With ClickFix Malware
Article
Threat Actors Leverage Fake JPEG Files for Initial Access
In a recent campaign, researchers observed threat actors using fake JPEG image files as a delivery mechanism to initiate the deployment of additional malicious components. The false JPEG files are typically distributed via phishing emails or other social engineering-based lures, and are actually PowerShell-based malware that deploys a trojanized version of ConnectWise ScreenConnect to establish and maintain persistence in the compromised environment.
What’s Notable and Unique
This campaign leverages JPEG images as the initial lure, where the images are not merely decoys but part of the infection workflow. Victims are typically led to download or open an image that triggers hidden execution logic or redirects them to a payload-delivery sequence that initiates later stages of the intrusion chain.
The attack chain is designed to blend into legitimate environments, making detection more difficult. Execution typically relies on scripted or native Windows components, often including PowerShell or other living-off-the-land binaries, enabling fileless or near-fileless execution and reducing forensic artifacts on disk.
The multistage design ensures that the initial JPEG does not directly contain the full payload but instead triggers retrieval or decryption steps that progressively assemble the final malicious components in memory.
Analyst Comments
This campaign illustrates how threat actors continue to blur the line between legitimate file handling and malicious execution chains, indicating potential overlap with remote management or administrative tooling. The use of JPEG-based staging combined with script-based execution reflects a broader evolution toward a stealth-first intrusion design, in which file formats serve as triggers rather than payload containers.
Sources
OPERATION SILENTCANVAS : JPEG BASED MULTISTAGE POWERSHELL INTRUSION
Podcast
Cyber Risk and Insurance for Law Firms
In this episode of Bytes of Insight, host Vinny Sakore is joined by Laura Zaroski, Managing Director of the Law Firms Group at Gallagher, as they discuss the evolution of cyber risk for law firms. Tune in for firsthand insights on how to select the right cyber policy, the incident response process, and the nuances of ransom payments and sensitive data.