EXPLORE

Article

Threat Actors Leveraging Gemini AI for All Attack Stages

Arete Analysis

State-backed threat actors are leveraging Google’s Gemini AI as a force multiplier to support all stages of the cyberattack lifecycle, from reconnaissance to post-compromise operations. According to the Google Threat Intelligence Group (GTIG), threat actors linked to the People’s Republic of China (PRC), Iran, North Korea, and other unattributed groups have misused Gemini to accelerate target profiling, synthesize open-source intelligence, identify official email addresses, map organizational structures, generate tailored phishing lures, translate content, conduct vulnerability testing, support coding tasks, and troubleshoot malware development. Cybercriminals are increasingly exploring AI-enabled tools and services to scale malicious activities, including social engineering campaigns such as ClickFix, demonstrating how generative AI is being integrated into both espionage and financially motivated threat operations. 

What’s Notable and Unique 

  • Threat actors are leveraging Gemini beyond basic reconnaissance, using it to generate polished, culturally nuanced phishing lures and sustain convincing multi-turn social engineering conversations that minimize traditional red flags.  

  • In addition, threat actors rely on Gemini for vulnerability research, malware debugging, code generation, command-and-control development, and technical troubleshooting, with PRC groups emphasizing automation and vulnerability analysis, Iranian actors focusing on social engineering and malware development, and North Korean actors prioritizing high-fidelity target profiling. 

  • Beyond direct operational support, adversaries have abused public generative AI platforms to host deceptive ClickFix instructions, tricking users into pasting malicious commands that deliver macOS variants of ATOMIC Stealer.  

  • AI is also being integrated directly into malware development workflows, as seen with CoinBait’s AI-assisted phishing kit capabilities and HonestCue’s use of the Gemini API to dynamically generate and execute in-memory C# payloads.  

  • Underground forums show strong demand for AI-powered offensive tools, with offerings like Xanthorox falsely marketed as custom AI but actually built on third-party commercial models integrated through open-source frameworks such as Crush, Hexstrike AI, LibreChat-AI, and Open WebUI, including Gemini. 

Analyst Comments 

The increasing misuse of generative AI platforms like Gemini highlights a rapidly evolving threat landscape in which state-backed and financially motivated actors leverage AI as a force multiplier for reconnaissance, phishing, malware development, and post-compromise operations. At the same time, large-scale model extraction attempts and API abuse demonstrate emerging risks to AI service integrity, intellectual property, and the broader AI-as-a-Service ecosystem. While these developments underscore the scalability and sophistication of AI-enabled threats, continued enforcement actions, strengthened safeguards, and proactive security testing by providers reflect ongoing efforts to mitigate abuse and adapt defenses in response to increasingly AI-driven adversaries. 

Sources 

Back to Blog Posts

Article

Critical MOVEit Automation Vulnerabilities Disclosed

A security advisory released by Progress Software details critical and high-severity vulnerabilities affecting their MOVEit Automation managed file transfer (MFT) solution. The vulnerabilities, tracked as CVE-2026-4670 and CVE-2026-5174, could allow a threat actor to bypass authentication and escalate privileges, leading to unauthorized access, administrative control, and data exposure. Cybercriminals have leveraged several MFT tools in previous campaigns, including the Accellion File Transfer Application (FTA), Fortra GoAnywhere MFT, and Cleo MFT. Flaws in MFT software are highly targeted by cybercriminals due to the volume and sensitivity of the data they control. 

What’s Notable and Unique

  • MOVEit Transfer was heavily exploited by the Cl0p ransomware group in the summer of 2023. While the window of exploit activity lasted only a few weeks, victim extortion and data leaks continued throughout the remainder of the year, leading to more than 70 class-action lawsuits filed in the U.S.

  • There is no workaround or hotfix for these vulnerabilities. To fully patch the flaws, MOVEit administrators need to perform a "full install" of the latest version, which will require taking the system offline.

  • Security researchers have discovered ~1,400 MOVEit Automation instances exposed to the internet, with dozens belonging to U.S. local and state government agencies.

Analyst Comments

While the vulnerabilities patched in Progress Software's recent release differ from the SQL injection vulnerability exploited by the Cl0p ransomware group in 2023, exploitation of CVE-2026-4670 and CVE-2026-5174 could lead to equally impactful outcomes. Beyond the immediate impacts on affected organizations, trusted data-exchange platforms provide threat actors with an avenue to obtain sensitive information and infect partner and supplier environments. Furthermore, Arete has seen the time window between disclosure and weaponization of critical vulnerabilities continue to shrink, especially as threat actors increasingly adopt AI-enabled tooling. As such, organizations should not only implement the patches released by Progress Software, but also hunt for typical post-compromise behavior like enumeration of the underlying database, the creation of new user accounts or users operating with unexpected administrator privileges, and the presence of unauthorized remote monitoring and management (RMM) tools. 

Sources

  • MOVEit Automation Critical Security Alert Bulletin – April 2026 – (CVE-2026-4670, CVE-2026-5174) 

  • From Breach to Courtroom: Inside the MOVEit Exploitation and Mass Litigation 

  • Progress warns of critical MOVEit Automation auth bypass flaw 

A graphic with futuristic lines showing a text saying Ransomware Trends and Data Insights, a monthly blog post.
A graphic with futuristic lines showing a text saying Ransomware Trends and Data Insights, a monthly blog post.

Article

Ransomware Trends & Data Insights: April 2026

The threat landscape has remained relatively predictable thus far in 2026. In April, Qilin dethroned Akira as the most active threat group for the month. Akira, who had been the top ransomware threat each month since July 2025, was still only slightly behind Qilin and had roughly the same activity level as in March. INC Ransom and DragonForce also remained active threats in April, with those four ransomware groups accounting for half of all ransomware and extortion activity observed by Arete.

A monthly graph showing the latest threat actor accounting for half of all ransomware and extortion acitivty observed by Arete.

Figure 1. Activity from the top 3 threat groups in April 2026

Throughout the month, analysts at Arete identified several trends behind the threat actors perpetrating cybercrime activities:

  • Multiple ransomware operations continue to leverage the Bring Your Own Vulnerable Driver (BYOVD) technique to disable endpoint security controls prior to ransomware deployment. Qilin has recently been observed leveraging a malicious file loaded via DLL side-loading along with vulnerable drivers, including rwdrv.sys and hlpdrv.sys, to gain kernel-level access and disable security processes. Arete observed Akira using the same vulnerable drivers in multiple engagements dating back to Q3 2025.


  • DragonForce has leveraged several of the same tools in recent engagements, including the remote desktop application Remotely Agent and the PoisonX.sys vulnerable driver. Additionally, open-source reporting indicates that the group recently used a Python-based backdoor known as VIPERTUNNEL to maintain reliable operator access and evade detection. DragonForce was responsible for over 7% of Arete ransomware engagements in April, and Arete notes increased activity from the group this year compared to 2025.


  • A social engineering tactic has reemerged in recent months in which threat actors impersonate IT and helpdesk staff via Microsoft Teams to contact employees and attempt to convince them to install remote access tools like Quick Assist, giving the threat actors remote access to the victim’s environment. This tactic was initially observed in late 2024 and early 2025 and was linked to now-defunct groups like Black Basta and Cactus, but has more recently been observed in intrusions linked to the Akira and Payouts King ransomware groups.

Sources

  • Arete Internal

Article

Payouts King Utilizes QEMU Emulator to Bypass EDR

Researchers recently identified threat actor campaigns leveraging QEMU, a free open-source virtual machine (VM) emulator, to evade endpoint security solutions. Since QEMU acts as a VM within the target environment, endpoint detection tools cannot scan inside the emulator or detect any malicious files or payloads QEMU contains. Although threat actors have been utilizing QEMU maliciously since 2020, recent activity is attributed to the Payouts King ransomware group and a cluster of threat actors believed to be initial access brokers who have also been exploiting the CitrixBleed2 vulnerability CVE-2025-5777.

What’s Notable and Unique

  • Payouts King has been observed deploying QEMU since November and uses the VM to create a reverse SSH backdoor to evade detection and install various tools, including Rclone, Chisel, and BusyBox.

  • In a separate campaign, threat actors are exploiting CVE-2025-5777, a Citrix NetScaler vulnerability that allows attackers to bypass authentication. Once they’ve gained initial access, the threat actors use QEMU to deploy tools inside the VM, which are then used to steal credentials, identify Kerberos usernames, perform Active Directory reconnaissance, and set up FTP servers for staging or data exfiltration.

Analyst Comments

Threat actors continue to focus their efforts on defense evasion, often leveraging legitimate, easily accessible tools such as QEMU. The continued use of QEMU by multiple threat actors highlights the effectiveness of these tactics and the difficulty in detecting and defending against them. To counter this campaign, organizations should proactively monitor for unauthorized QEMU installations, abnormal scheduled tasks, and port forwarding rules. 

 Sources

  • QEMU abused to evade detection and enable ransomware delivery

Article

Microsoft Teams Continues to be Leveraged in Social Engineering Attacks

Microsoft warns that threat actors are increasingly abusing Microsoft Teams and relying on legitimate tools to gain access and conduct lateral movement within enterprise networks. The threat actors impersonate IT or helpdesk staff to contact employees via cross-tenant chats and trick them into granting remote access for data theft. Microsoft has observed multiple intrusions with a similar attack chain that utilized commercial remote management software, like Quick Assist and the Rclone utility, to transfer files to an external cloud storage service. This tactic, notably associated with Black Basta and Cactus ransomware operations in late 2024 and early 2025, appears to have resurfaced, with similar activity more recently observed in intrusions linked to the Akira and Payouts King ransomware groups.

What’s Notable and Unique

  • Initial access is achieved by leveraging external collaboration features in Microsoft Teams to allow impersonation of internal support personnel, tricking users into bypassing security warnings. This reflects abuse of legitimate functionality rather than exploitation of a Microsoft Teams vulnerability.


  • Following initial access, attackers conduct rapid reconnaissance using Command Prompt and PowerShell to assess privileges, domain membership, and opportunities for lateral movement. Persistence is maintained through Windows Registry modifications, after which attackers leveraged WinRM for lateral movement, targeting domain-joined systems and high-value assets, including domain controllers.


  • Malicious payloads were staged in user-writable directories and executed through DLL side-loading via trusted, signed applications, enabling covert code execution while blending with legitimate activity. Additional remote management tools were also deployed to support broader access, while Rclone or similar utilities were used to stage and exfiltrate sensitive data to external cloud storage. 

Analyst Comments

This activity highlights how modern threat actors can leverage trusted collaboration workflows, remote management tools, and stealthy exfiltration techniques to conduct intrusions through a combination of social engineering and misuse of legitimate functionality. Effective defense depends on layered mitigations that combine identity controls, restricted remote administration, endpoint hardening, network protections, and user awareness measures to disrupt attacker activity at multiple stages of the intrusion lifecycle. To mitigate the risk of this and similar campaigns, users should treat external Teams contacts as untrusted by default, and administrators should restrict or closely monitor remote assistance tools while limiting WinRM usage to controlled systems. 

Sources

  • Cross‑tenant helpdesk impersonation to data exfiltration: A human-operated intrusion playbook

  • Microsoft: Teams increasingly abused in helpdesk impersonation attacks

  • Payouts King Takes Aim at the Ransomware Throne