Blog
/
AI
/
March 7, 2024

Defending Against the New Normal in Cybercrime: AI

This blog outlines research & data points on the evolving threat landscape, the impact of malicious AI, and why proactive cyber readiness is essential.
Inside the SOC
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
Written by
Max Heinemeyer
Global Field CISO
Default blog imageDefault blog imageDefault blog imageDefault blog imageDefault blog imageDefault blog image
07
Mar 2024

AI in Cyber Security

Over the last 18 months, discussions about artificial intelligence (AI) – specifically generative AI – ranged from excitement and optimism about its transformative potential to fear and uncertainty about the new risks it introduces.  

New research1 commissioned by Darktrace shows that 89 percent of IT security teams polled globally believe AI-augmented cyber threats will have a significant impact on their organization within the next two years, yet 60 percent believe they are currently unprepared to defend against these attacks. Their concerns include increased volume and sophistication of malware that targets known vulnerabilities and increased exposure of sensitive or proprietary information from using generative AI tools.  

At Darktrace, we monitor trends across our global customer base to understand how the challenges facing security teams are evolving alongside industry advancements in AI. We’ve observed that AI, automation, and cybercrime-as-a-service have increased the speed, sophistication and efficacy of cyber security attacks.  

How AI Impacts Phishing Attempts

Darktrace has observed immediate impacts on phishing, which remains one of the most common forms of attack. In April 2023, Darktrace shared research that found a 135 percent increase in ‘novel social engineering attacks’ in the first two months of 2023, corresponding with the widespread adoption of ChatGPT2. These phishing attacks showed a strong linguistic deviation – semantically and syntactically – compared to other phishing emails, which suggested to us that generative AI is providing an avenue for threat actors to craft sophisticated and targeted attacks at speed and scale. A year later, we’ve seen this trend continue. Darktrace customers received approximately 2,867,000 phishing emails in December 2023 alone, a 14 percent increase on what was observed months prior in September3. Between September and December 2023, phishing attacks that used novel social engineering techniques grew by 35 percent on average across the Darktrace customer base4.  

These observations reinforce trends that others in the industry have shared. For example, Microsoft and OpenAI recently published research on tactics, techniques, and procedures (TTPs) augmented by large language models (LLMs) that they have observed nation-state threat actors using. That includes using LLMs to draft and generate social engineering attacks, inform reconnaissance, assist with vulnerability research and more.  

The Rise of Cybercrime-as-as-a-Service

The increasing cyber challenge facing defenders cannot be attributed to AI alone. The rise of cybercrime as-a-service is also changing the dynamic. Darktrace’s 2023 End of Year Threat Report found that cybercrime-as-a-service continue to dominate the threat landscape, with malware-as-a-Service (MaaS) and ransomware-as-a-Service (RaaS) tools making up most malicious tools in use by attackers. The as-a-Service ecosystem can provide attackers with everything from pre-made malware to templates for phishing emails, payment processing systems and even helplines to enable bad actors to mount attacks with limited technical knowledge.  

These trends make it clear that attackers now have a more widely accessible toolbox that reduces their barriers.

AI Enabling Accidental Insider Threats

However, the new risks facing businesses aren’t from external threat actors alone. Use of generative AI tools within the enterprise introduces a new category of accidental insider threats. Employees using generative AI tools now have easier access to more organizational data than ever before. Even the most well-intentioned employee could unintentionally leak or access restricted, sensitive data via these tools. In the second half of 2023, we observed that approximately half of Darktrace customers had employees accessing generative AI services. As this continues to increase, organizations need policies in place to guide the use cases for generative AI tools as well as strong data governance and the ability to enforce these policies to minimize risk.  

It is inevitable that AI will increase the risks and threats facing an organization, but this is not an unsolvable challenge from a defensive perspective. While advancements in generative AI may be worsening issues like novel social engineering and creating new types of accidental insider threats, AI itself offers a strong defense.  

The Shift to Proactive Cyber Readiness

According to the World Economic Forum’s Global Cybersecurity Outlook 2024, the number of organizations that “maintain minimum viable cyber resilience is down 30 percent compared to 2023”, and “while large organizations have demonstrated gains in cyber resilience, small and medium-sized companies showed significant decline.” The importance of cyber resilience cannot be understated in the face of today’s increasingly as-a-service, automated, and AI-augmented threat landscape.  

Historically, organizations wait for incidents to happen and rely on known attack data for threat detection and response, making it nearly impossible to identify never-before-seen threats. The traditional security stack has also relied heavily on point solutions focused on protecting different pieces of the digital environment, with individual tools for endpoint, email, network, on-premises data centers, SaaS applications, cloud, OT and beyond. These point solutions fail to correlate disparate incidents to form a complete picture of an orchestrated attack. Even with the addition of tools that can stitch together events from across the enterprise, they are in a reactive state that focuses heavily on threat detection and response.  

Organizations need to evolve from a reactive posture to a stance of proactive cyber readiness. To do so, they need an approach that proactively identifies internal and external vulnerabilities, identifies gaps in security policy and process before an attack occurs, breaks down silos to investigate all threats (known and unknown) during an attack, and uplifts the human analyst beyond menial tasks to incident validation and recovery after an attack.  

AI can help break down silos within the SOC and provide a more proactive approach to scale up and augment defenders. It provides richer context when it is fed information from multiple systems, data sets, and tools within the stack and can build an in-depth, real-time behavioural understanding of a business that humans alone cannot.

Lessons From AI in the SOC

At Darktrace, we’ve been applying AI to the challenge of cyber security for more than ten years, and we know that proactive cyber readiness requires the right mix of people, process, and technology.  

When the right AI is applied responsibly to the right cyber security challenge, the impact on both the human security team and the business is profound.

AI can bring machine speed and scale to some of the most time-intensive, error-prone, and psychologically draining components of cyber security, helping humans focus on the value-added work that only they can provide. Incident response and continuous monitoring are two areas where AI has already been proven to effectively augment defenders. For example, a civil engineering company used Darktrace’s AI to uplift its SOC team from the repetitive, manual tasks of analyzing and responding to email incidents. The analysts estimated they were each spending 10 hours per week on email incident analysis. With AI autonomously analyzing and responding to email incidents, the analysts could gain approximately 20 percent of their time back to focus on proactive cyber security measures

An effective human-AI partnership is key to proactive cyber readiness and can directly benefit the work-life of defenders. It can help to reduce burnout, support data-driven decision-making, and reduce the reliance on hard-to-find, specialized talent that has created a skills shortage in cyber security for many years. Most importantly, AI can free up team members to focus on more meaningful tasks, such as compliance initiatives, user education, and sophisticated threat hunting.  

Advancements in AI are happening at a rapid pace. As we’ve already observed, attackers will be watching these developments and looking for ways to use it to their advantage. Luckily, AI has already proved to be an asset for defenders, and embracing a proactive approach to cyber resilience can help organizations increase their readiness for this next phase. Prioritizing cyber security will be an enabler of innovation and progress as AI development continues.  

--

Join Darktrace on 9 April for a virtual event to explore the latest innovations needed to get ahead of the rapidly evolving threat landscape. Register today to hear more about our latest innovations coming to Darktrace’s offerings.

References

[1] The survey was undertaken by AimPoint Group & Dynata on behalf Darktrace between December 2023 & January 2024. The research polled 1773 security professionals in positions across the security team from junior roles to CISOs, across 14 countries – Australia, Brazil, France, Germany, Italy, Japan, Mexico, Netherlands, Singapore, Spain, Sweden, UAE, UK, and USA.

[2] Based on the average change in email attacks between January and February 2023 detected across Darktrace/Email deployments with control of outliers.

[3] Average calculated across Darktrace customers from 31st August to 21st December.

[4] Average calculated across Darktrace customers from 31st August to 21st December. Novel social engineering attacks use linguistic techniques that are different to techniques used in the past, as measured by a combination of semantics, phrasing, text volume, punctuation, and sentence length.

Inside the SOC
Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.
Written by
Max Heinemeyer
Global Field CISO

More in this series

No items found.

Blog

/

Email

/

May 29, 2025

Why attack-centric approaches to email security can’t cope with modern threats

Default blog imageDefault blog image

What’s the problem with an attack-centric mindset?

For decades, traditional email security strategies have been built around an attack-centric mindset. Secure Email Gateways (SEGs) and other legacy solutions operate on the principle of identifying and blocking known threats. These systems rely heavily on predefined threat intelligence – blacklists, malware signatures, and reputation-based analysis – to filter out malicious content before it reaches the inbox.

While this approach was sufficient when email threats were relatively static and signature-based, it’s increasingly ineffective against the sophistication of modern attacks. Techniques like spear phishing, business email compromise (BEC), and supply chain attacks often bypass traditional SEG defenses because they lack obvious malicious indicators. Instead, they leverage social engineering, look-alike domains, and finely tuned spoofing tactics that are designed to evade detection.

The challenge extends beyond just legacy SEGs. Many modern email security providers have inherited the same attack-centric principles, even if they've reimagined the technology stack. While some vendors have shifted to API-based deployments and incorporated AI to automate pattern recognition, the underlying approach remains the same: hunting for threats based on known indicators. This methodology, though it’s undergone modernization using AI, still leaves gaps when it comes to novel, hyper-targeted threats that manipulate user behavior rather than deploy predictable malicious signatures. Attack-centric security will always remain one step behind the attacker.

By the way, native email security already covers the basics

One of the most overlooked realities in email security is that native solutions like Microsoft 365’s built-in security already handle much of the foundational work of attack-centric protection. Through advanced threat intelligence, anti-phishing measures, and malware detection, Microsoft 365 actively scans incoming emails for known threats, using global telemetry to identify patterns and block suspicious content before it even reaches the user’s inbox.

This means that for many organizations, a baseline level of protection against more obvious, signature-based attacks is already in place – but many are still disabling these protections in favour of another attack-centric solution. By layering another attack-centric solution on top, they are effectively duplicating efforts without enhancing their security posture. This overlap can lead to unnecessary complexity, higher costs, and a false sense of enhanced protection when in reality, it’s more of the same.

Rather than duplicating attack-centric protections, the real opportunity lies in addressing the gaps that remain: the threats that are specifically crafted to evade traditional detection methods. This is where a business-centric approach becomes indispensable, complementing the foundational security that’s already built into your infrastructure.

Introducing… the business-centric approach

To effectively defend against advanced threats, organizations need to adopt a business-centric approach to email security. Unlike attack-centric models that hunt for known threats, business-centric security focuses on understanding the typical behaviors, relationships, and communication patterns within your organization. Rather than solely reacting to threats as they are identified, this model continuously learns what “normal” looks like for each user and each inbox.

By establishing a baseline of expected behaviors, business-centric solutions can rapidly detect anomalies that suggest compromise, such as sudden changes in sending patterns, unusual login locations, or subtle shifts in communication tone. This proactive detection method is especially powerful against spear phishing, business email compromise (BEC), and supply chain attacks that are engineered to bypass static defenses. This approach also scales with your organization, learning and adapting as new users are onboarded, communication patterns evolve, and external partners are added.

In an era where AI-driven threats are becoming the norm, having email security that knows your users and inboxes better than the attacker does is a critical advantage.

Why native + business-centric email security is the winning formula

By pairing native security with a business-centric model, organizations can cover the full spectrum of threats – from signature-based malware to sophisticated, socially engineered attacks. Microsoft 365’s in-built security manages the foundational risks, while business-centric defense identifies subtle anomalies and targeted threats that legacy approaches miss.

Layering Darktrace on top of your native Microsoft security eliminates duplicate capabilities, costs and workflows without reducing functionality

Rather than layering redundant attack-centric solutions on top of existing protections, the future of email security lies in leveraging what’s already in place and building on it with smarter, behavior-based detection. The Swiss Cheese Model is a useful one to refer to here: by acknowledging that no single defense can offer complete protection, layering defenses that plug each other’s gaps – like slices of Swiss cheese – becomes critical.

This combination also allows security teams to focus their efforts more effectively. With native solutions catching broad-based, known threats, the business-centric layer can prioritize real anomalies, minimizing false positives and accelerating response times. Organizations benefit from reduced overlap, streamlined costs, and a stronger overall security posture.

Download the full guide to take the first step towards achieving your next-generation security stack.

[related-resource]

Continue reading
About the author
Carlos Gray
Senior Product Marketing Manager, Email

Blog

/

Network

/

May 30, 2025

PumaBot: Novel Botnet Targeting IoT Surveillance Devices

password login screen on computerDefault blog imageDefault blog image

Introduction: PumaBot attacking IoT devices

Darktrace researchers have identified a custom Go-based Linux botnet named “PumaBot” targeting embedded Linux Internet of Things (IoT) devices. Rather than scanning the Internet, the malware retrieves a list of targets from a command-and-control (C2) server and attempts to brute-force SSH credentials. Upon gaining access, it receives remote commands and establishes persistence using system service files. This blog post provides a breakdown of its key functionalities, and explores binaries related to the campaign.

Technical Analysis

Filename: jierui

md5: cab6f908f4dedcdaedcdd07fdc0a8e38

The Go-based botnet gains initial access through brute-forcing SSH credentials across a list of harvested IP addresses. Once it identifies a valid credential pair, it logs in, deploys itself, and begins its replication process.

Overview of Jierui functions
Figure 1: Overview of Jierui functions.

The domain associated with the C2 server did not resolve to an IP address at the time of analysis. The following details are a result of static analysis of the malware.

The malware begins by retrieving a list of IP addresses of likely devices with open SSH ports from the C2 server (ssh.ddos-cc[.]org) via the getIPs() function. It then performs brute-force login attempts on port 22 using credential pairs also obtained from the C2 through the readLinesFromURL(), brute(), and trySSHLogin() functions.

Within trySSHLogin(), the malware performs several environment fingerprinting checks. These are used to avoid honeypots and unsuitable execution environments, such as restricted shells. Notably, the malware checks for the presence of the string “Pumatronix”, a manufacturer of surveillance and traffic camera systems, suggesting potential IoT targeting or an effort to evade specific devices [1].

Fingerprinting of “Pumatronix”.
Figure 2: Fingerprinting of “Pumatronix”.

If the environment passes these checks, the malware executes uname -a to collect basic system information, including the OS name, kernel version, and architecture. This data, along with the victim's IP address, port, username, and password, is then reported back to the C2 in a JSON payload.

Of note, the bot uses X-API-KEY: jieruidashabi, within a custom header when it communicates with the C2 server over HTTP.

The malware writes itself to /lib/redis, attempting to disguise itself as a legitimate Redis system file. It then creates a persistent systemd service in /etc/systemd/system, named either redis.service or mysqI.service (note the spelling of mysql with a capital I) depending on what has been hardcoded into the malware. This allows the malware to persist across reboots while appearing benign.

[Unit]
Description=redis Server Service

[Service]
Type=simple
Restart=always
RestartSec=1
User=root
ExecStart=/lib/redis e

[Install]
WantedBy=multi-user.target

In addition to gaining persistence with a systemd service, the malware also adds its own SSH keys into the users’ authorized_keys file. This ensures that access can be maintained, even if the service is removed.

A function named cleankill() contains an infinite loop that repeatedly attempts to execute the commands “xmrig” and “networkxm”. These are launched without full paths, relying on the system's PATH variable suggesting that the binaries may be downloaded or unpacked elsewhere on the system. The use of “time.Sleep” between attempts indicates this loop is designed to ensure persistence and possibly restart mining components if they are killed or missing.

During analysis of the botnet, Darktrace discovered related binaries that appear to be part of a wider campaign targeting Linux systems.

Filename: ddaemon
Md5: 48ee40c40fa320d5d5f8fc0359aa96f3

Ddaemon is a Go-based backdoor. The malware begins by parsing command line arguments and if conditions are met, enters a loop where it periodically verifies the MD5 hash of the binary. If the check fails or an update is available, it downloads a new version from a C2 server (db.17kp[.]xyz/getDdaemonMd5), verifies it and replaces the existing binary with a file of the same name and similar functionality (8b37d3a479d1921580981f325f13780c).

The malware uses main_downloadNetwork() to retrieve the binary “networkxm” into /usr/src/bao/networkxm. Additionally, the bash script “installx.sh” is also retrieved from the C2 and executed. The binary ensures persistence by writing a custom systemd service unit that auto starts on boot and executes ddaemon.

Filename: networkxm
Md5: be83729e943d8d0a35665f55358bdf88

The networkxm binary functions as an SSH brute-force tool, similar to the botnet. First it checks its own integrity using MD5 hashes and contacts the C2 server (db.17kp[.]xyz) to compare its hash with the latest version. If an update is found, it downloads and replaces itself.

Part of networkxm checking MD5 hash.
Figure 3: Part of networkxm checking MD5 hash.
MD5 hash
Figure 4: MD5 hash

After verifying its validity, it enters an infinite loop where it fetches a password list from the C2 (/getPassword), then attempts SSH connections across a list of target IPs from the /getIP endpoint. As with the other observed binaries, a systemd service is created if it doesn’t already exist for persistence in /etc/systemd/system/networkxm.service.

Bash script installx.sh.
Figure 5: Bash script installx.sh.

Installx.sh is a simple bash script used to retrieve the script “jc.sh” from 1.lusyn[.]xyz, set permissions, execute and clear bash history.

Figure 6: Snippet of bash script jc.sh.

The script jc.sh starts by detecting the operating system type Debian-based or Red Hat-based and determines the location of the pam_unix.so file. Linux Pluggable Authentication Modules (PAM) is a framework that allows for flexible and centralized user authentication on Linux systems. PAM allows system administrators to configure how users are authenticated for services like login, SSH, or sudo by plugging in various authentication modules.

Jc.sh then attempts to fetch the current version of PAM installed on the system and formats that version to construct a URL. Using either curl or wget, the script downloads a replacement pam_unix.so file from a remote server and replaces the existing one, after disabling file immutability and backing up the original.

The script also downloads and executes an additional binary named “1” from the same remote server. Security settings are modified including enabling PAM in the SSH configuration and disabling SELinux enforcement, before restarting the SSH service. Finally, the script removes itself from the system.

Filename: Pam_unix.so_v131
md5: 1bd6bcd480463b6137179bc703f49545

Based on the PAM version that is retrieved from the bash query, the new malicious PAM replaces the existing PAM file. In this instance, pam_unix.so_v131 was retrieved from the server based on version 1.3.1. The purpose of this binary is to act as a rootkit that steals credentials by intercepting successful logins. Login data can include all accounts authenticated by PAM, local and remote (SSH). The malware retrieves the logged in user, the password and verifies that the password is valid. The details are stored in a file “con.txt” in /usr/bin/.

Function storing logins to con.txt
Figure 7: Function storing logins to con.txt

Filename: 1

md5: cb4011921894195bcffcdf4edce97135

In addition to the malicious PAM file, a binary named “1” is also retrieved from the server http://dasfsdfsdfsdfasfgbczxxc[.]lusyn[.]xyz/jc/1. The binary “1” is used as a watcher for the malicious PAM file using inotify to monitor for “con.txt” being written or moved to /usr/bin/.

Following the daemonize() function, the binary is run daemonized ensuring it runs silently in the background. The function read_and_send_files() is called which reads the contents of “/usr/bin/con.txt”, queries the system IP with ifconfig.me, queries SSH ports and sends the data to the remote C2 (http://dasfsdfsdfsdfasfgbczxxc[.]lusyn[.]xyz/api/).

Command querying SSH ports.
Figure 8: Command querying SSH ports.

For persistence, a systemd service (my_daemon.service) is created to autostart the binary and ensure it restarts if the service has been terminated. Finally, con.txt is deleted, presumably to remove traces of the malware.

Conclusion

The botnet represents a persistent Go-based SSH threat that leverages automation, credential brute-forcing, and native Linux tools to gain and maintain control over compromised systems. By mimicking legitimate binaries (e.g., Redis), abusing systemd for persistence, and embedding fingerprinting logic to avoid detection in honeypots or restricted environments, it demonstrates an intent to evade defenses.

While it does not appear to propagate automatically like a traditional worm, it does maintain worm-like behavior by brute-forcing targets, suggesting a semi-automated botnet campaign focused on device compromise and long-term access.

[related-resource]

Recommendations

  1. Monitor for anomalous SSH login activity, especially failed login attempts across a wide IP range, which may indicate brute-force attempts.
  2. Audit systemd services regularly. Look for suspicious entries in /etc/systemd/system/ (e.g., misspelled or duplicate services like mysqI.service) and binaries placed in non-standard locations such as /lib/redis.
  3. Inspect authorized_keys files across user accounts for unknown SSH keys that may enable unauthorized access.
  4. Filter or alert on outbound HTTP requests with non-standard headers, such as X-API-KEY: jieruidashabi, which may indicate botnet C2 communication.
  5. Apply strict firewall rules to limit SSH exposure rather than exposing port 22 to the internet.

Appendices

References

1.     https://pumatronix.com/

Indicators of Compromise (IoCs)

Hashes

cab6f908f4dedcdaedcdd07fdc0a8e38 - jierui

a9412371dc9247aa50ab3a9425b3e8ba - bao

0e455e06315b9184d2e64dd220491f7e - networkxm

cb4011921894195bcffcdf4edce97135 - 1
48ee40c40fa320d5d5f8fc0359aa96f3 - ddaemon
1bd6bcd480463b6137179bc703f49545 - pam_unix.so_v131

RSA Key

ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQC0tH30Li6Gduh0Jq5A5dO5rkWTsQlFttoWzPFnGnuGmuF+fwIfYvQN1z+WymKQmX0ogZdy/CEkki3swrkq29K/xsyQQclNm8+xgI8BJdEgTVDHqcvDyJv5D97cU7Bg1OL5ZsGLBwPjTo9huPE8TAkxCwOGBvWIKUE3SLZW3ap4ciR9m4ueQc7EmijPHy5qds/Fls+XN8uZWuz1e7mzTs0Pv1x2CtjWMR/NF7lQhdi4ek4ZAzj9t/2aRvLuNFlH+BQx+1kw+xzf2q74oBlGEoWVZP55bBicQ8tbBKSN03CZ/QF+JU81Ifb9hy2irBxZOkyLN20oSmWaMJIpBIsh4Pe9 @root

Network

http://ssh[.]ddos-cc.org:55554

http://ssh[.]ddos-cc.org:55554/log_success

http://ssh[.]ddos-cc.org:55554/get_cmd

http://ssh[.]ddos-cc.org:55554/pwd.txt

https://dow[.]17kp.xyz/

https://input[.]17kp.xyz/

https://db[.]17kp[.]xyz/

http://1[.]lusyn[.]xyz

http://1[.]lusyn[.]xyz/jc/1

http://1[.]lusyn[.]xyz/jc/jc.sh

http://1[.]lusyn[.]xyz/jc/aa

http://1[.]lusyn[.]xyz/jc/cs

http://dasfsdfsdfsdfasfgbczxxc[.]lusyn[.]xyz/api

http://dasfsdfsdfsdfasfgbczxxc[.]lusyn[.]xyz/jc

Detection Rule

rule Linux_PumaBot

{

  meta:

      description = "Rule to match on PumaBot samples"

      author = "tgould@cadosecurity.com"

  strings:

      $xapikey = "X-API-KEY" ascii

      $get_ips = "?count=5000" ascii

      $exec_start = "ExecStart=/lib/redis" ascii

      $svc_name1 = "redis.service" ascii

      $svc_name2 = "mysqI.service" ascii

      $uname = "uname -a" ascii

      $pumatronix = "Pumatronix" ascii

  condition:

      uint32(0) == 0x464c457f and

      all of (

          $xapikey,

          $uname,

          $get_ips,

          $exec_start

      ) and any of (

          $svc_name1,

          $svc_name2

      ) and $pumatronix

}

Continue reading
About the author
Tara Gould
Threat Researcher
Your data. Our AI.
Elevate your network security with Darktrace AI