The rapid adoption of large language models (LLMs) has fueled speculation that cybercriminals may utilize LLMs to improve and automate their attacks.However, so far, the security community has had only anecdotal evidence of attackers using LLMs, lacking large-scale data on the extent of real-world malicious LLM usage.In this joint work between academic researchers and Barracuda Networks, we present the first large-scale study measuring AIgenerated attacks in-the-wild.In particular, we focus on the use of LLMs by attackers to craft the text of malicious emails by analyzing a corpus of hundreds of thousands of real-world malicious emails detected by Barracuda.The key challenge in this analysis is determining ground truth: we cannot know for certain whether an email is LLM or human-generated.To overcome this challenge, we observe that, prior to the launch of ChatGPT, email text was almost certainly not LLM-generated.Armed with this insight, we run three state-of-the-art LLM detection methods on our corpus and calibrate them against pre-ChatGPT emails, as well as against a diverse set of LLM-generated emails we create ourselves.Since the launch of ChatGPT, all three detection methods indicate that attackers have steadily increased their use of LLMs to * Work done at Columbia University.

This paper empirically evaluates the efficacy of two ubiquitous forms of enterprise security training: annual cybersecurity awareness training and embedded anti-phishing training exercises. Specifically, our work analyzes the results of an 8-month randomized controlled experiment involving ten simulated phishing campaigns sent to over 19,500 employees at a large healthcare organization. Our results suggest that these efforts offer limited value. First, we find no significant relationship between whether users have recently completed cybersecurity awareness training and their likelihood of failing a phishing simulation. Second, when evaluating recipients of embedded phishing training, we find that the absolute difference in failure rates between trained and untrained users is extremely low across a variety of training content. Third, we observe that most users spend minimal time interacting with embedded phishing training material in-the-wild; and that for specific types of training content, users who receive and complete more instances of the training can have an increased likelihood of failing subsequent phishing simulations. Taken together, our results suggest that anti-phishing training programs, in their current and commonly deployed forms, are unlikely to offer significant practical value in reducing phishing risks.

Email service has increasingly been outsourced to cloud-based providers and so too has the task of filtering such messages for potential threats. Thus, customers will commonly direct that their incoming email is first sent to a third-party email filtering service (e.g., Proofpoint or Barracuda) and only the "clean" messages are then sent on to their email hosting provider (e.g., Gmail or Microsoft Exchange Online). However, this loosely coupled approach can, in theory, be bypassed if the email hosting provider is not configured to only accept messages that arrive from the email filtering service. In this paper we demonstrate that such bypasses are commonly possible. We document a multi-step methodology to infer if an organization has correctly configured its email hosting provider to guard against such scenarios. Then, using an empirical measurement of edu and com domains as a case study, we show that 80% of such organizations making use of popular cloud-based email filtering services can be bypassed in this manner. We also discuss reasons that lead to such misconfigurations and outline challenges in hardening the binding between email filtering and hosting providers.

Enterprise-scale mandatory password changes are disruptive, complex endeavors that require the entire workforce to prioritize a goal that is often secondary to most users. While ample literature exists around user perceptions and struggles, there are few “best practices” from the perspective of the enterprise—either to achieve the end goal or to minimize IT costs. In this paper, we provide an empirical analysis of an enterprise-scale mandatory password change, covering almost 10,000 faculty and staff at an academic institution. Using a combination of user notifications logs, password update records, and help desk ticket information, we construct an empirical model of user response over time. In particular, we characterize the elements of the campaign that relate to ideal and non-ideal outcomes, including unnecessary user actions and IT help desk overhead. We aim to provide insight into successes and challenges that can generalize to other organizations implementing similar initiatives.

The critical role played by email has led to a range of extension protocols (e.g., SPF, DKIM, DMARC) designed to protect against the spoofing of email sender domains. These protocols are complex as is, but are further complicated by automated email forwarding — used by individual users to manage multiple accounts and by mailing lists to redistribute messages. In this paper, we explore how such email forwarding and its implementations can break the implicit assumptions in widely deployed anti-spoofing protocols. Using large-scale empirical measurements of 20 email forwarding services (16 leading email providers and four popular mailing list services), we identify a range of security issues rooted in forwarding behavior and show how they can be combined to reliably evade existing anti-spoofing controls. We further show how these issues allow attackers to not only deliver spoofed email messages to prominent email providers (e.g., Gmail, Microsoft Outlook, and Zoho), but also reliably spoof email on behalf of tens of thousands of popular domains including sensitive domains used by organizations in government (e.g., state.gov), finance (e.g., transunion.com), law (e.g., perkinscoie.com) and news (e.g., washingtonpost.com) among others.

Consumer mobile spyware apps covertly monitor a user's activities (i.e., text messages, phone calls, e-mail, location, etc.) and transmit that information over the Internet to support remote surveillance. Unlike conceptually similar apps used for state espionage, so-called "stalkerware" apps are mass-marketed to consumers on a retail basis and expose a far broader range of victims to invasive monitoring. Today the market for such apps is large enough to support dozens of competitors, with individual vendors reportedly monitoring hundreds of thousands of phones. However, while the research community is well aware of the existence of such apps, our understanding of the mechanisms they use to operate remains ad hoc. In this work, we perform an in-depth technical analysis of 14 distinct leading mobile spyware apps targeting Android phones. We document the range of mechanisms used to monitor user activity of various kinds (e.g., photos, text messages, live microphone access) — primarily through the creative abuse of Android APIs. We also discover previously undocumented methods these apps use to hide from detection and to achieve persistence. Additionally, we document the measures taken by each app to protect the privacy of the sensitive data they collect, identifying a range of failings on the part of spyware vendors (including privacy-sensitive data sent in the clear or stored in the cloud with little or no protection).

We examine the security of home smart locks: cyber-physical devices that replace traditional door locks with deadbolts that can be electronically controlled by mobile devices or the lock manufacturer's remote servers. We present two categories of attacks against smart locks and analyze the security of five commercially-available locks with respect to these attacks. Our security analysis reveals that flaws in the design, implementation, and interaction models of existing locks can be exploited by several classes of adversaries, allowing them to learn private information about users and gain unauthorized home access. To guide future development of smart locks and similar Internet of Things devices, we propose several defenses that mitigate the attacks we present. One of these defenses is a novel approach to securely and usably communicate a user's intended actions to smart locks, which we prototype and evaluate. Ultimately, our work takes a first step towards illuminating security challenges in the system design and novel functionality introduced by emerging IoT systems.

As miscreants routinely hijack thousands of vulnerable web servers weekly for cheap hosting and traffic acquisition, security services have turned to notifications both to alert webmasters of ongoing incidents as well as to expedite recovery. In this work we present the first large-scale measurement study on the effectiveness of combinations of browser, search, and direct webmaster notifications at reducing the duration a site remains compromised. Our study captures the life cycle of 760,935 hijacking incidents from July, 2014--June, 2015, as identified by Google Safe Browsing and Search Quality. We observe that direct communication with webmasters increases the likelihood of cleanup by over 50% and reduces infection lengths by at least 62%. Absent this open channel for communication, we find browser interstitials---while intended to alert visitors to potentially harmful content---correlate with faster remediation. As part of our study, we also explore whether webmasters exhibit the necessary technical expertise to address hijacking incidents. Based on appeal logs where webmasters alert Google that their site is no longer compromised, we find 80% of operators successfully clean up symptoms on their first appeal. However, a sizeable fraction of site owners do not address the root cause of compromise, with over 12% of sites falling victim to a new attack within 30 days. We distill these findings into a set of recommendations for improving web security and best practices for webmasters.

Today, web injection manifests in many forms, but fundamentally occurs when malicious and unwanted actors tamper directly with browser sessions for their own profit. In this work we illuminate the scope and negative impact of one of these forms, ad injection, in which users have ads imposed on them in addition to, or different from, those that websites originally sent them. We develop a multi-staged pipeline that identifies ad injection in the wild and captures its distribution and revenue chains. We find that ad injection has entrenched itself as a cross-browser monetization platform impacting more than 5% of unique daily IP addresses accessing Google -- tens of millions of users around the globe. Injected ads arrive on a client's machine through multiple vectors: our measurements identify 50,870 Chrome extensions and 34,407 Windows binaries, 38% and 17% of which are explicitly malicious. A small number of software developers support the vast majority of these injectors who in turn syndicate from the larger ad ecosystem. We have contacted the Chrome Web Store and the advertisers targeted by ad injectors to alert each of the deceptive practices involved.