Online scams continue to cause substantial financial and personal harm. As a result, detection systems based on Large Language Models (LLMs) have been integrated into security products ranging from email gateways and browser extensions to fraud-monitoring dashboards. As this adoption accelerates, a common belief has taken hold: that these models are broadly suitable for scam detection. In this work, we investigate whether LLMs, with their strong capabilities in understanding intent, context, and reasoning, can effectively detect scams across diverse scenarios without task-specific fine-tuning. We curate and release a unique benchmark dataset of real-world scams spanning multiple formats and topics. We evaluate nine LLMs of varying sizes and architectures, examining their performance under different prompting strategies and comparing them to a fine-tuned BERT-based classifier. Our results show that while larger LLMs generally outperform smaller ones, effective prompting substantially boosts the performance of smaller models. Moreover, LLMs are better at generalizing to unseen scams compared to fine-tuned models, suggesting that pre-trained knowledge contributes meaningfully to scam detection. We release our dataset and evaluation framework to facilitate future research in robust scam detection using language models.

The Better Business Bureau ranked online e-commerce scams as the top consumer threat for 2022. Our measurements of real consumer devices confirm that e-commerce scams receive large traffic volumes, a total of 6.3M visits during seven months. In this work, we study e-commerce scams in depth and design a detection classifier that combines novel features that target salient characteristics of e-commerce scam websites and features for detecting malicious and scam domains proposed by prior work. In addition, we specify a method for automatically creating reliable ground-truth sets that are an order of magnitude larger than that of prior work. We use this data set to evaluate the classifier and achieve a high 0.973 F1-score (Prec: 0.988, Rec: 0.959). In a best-effort comparison, we demonstrate that our classifier outperforms the F-1 score of the prior art by 11% and that our novel features offer an F1-score boost of 4.3% over the features used in the prior art. In addition, we deploy our classifier in a real-world setting, analyze over 760K e-shops visited by real users, and identify 10% of those as e-commerce scams. We demonstrate that the classifier has a low False Positive rate in real-world settings and can protect over 176K users in one week.

Trust in Secure Sockets Layer–based communications is traditionally provided by Certificate (or Certification) Authorities (CAs) in the form of signed certificates. Checking the validity of a certificate involves three steps: (i) checking its expiration date, (ii) verifying its signature, and (iii) ensuring that it is not revoked. Currently, such certificate revocation checks (i.e., step (iii) above) are done either via Certificate Revocation Lists (CRLs), or Online Certificate Status Protocol (OCSP) servers. Unfortunately, despite the existence of these revocation checks, sophisticated cyber-attackers can still trick web browsers to trust a revoked certificate, believing that it is still valid. Although frequently updated , nonced , and timestamped certificates can reduce the frequency and impact of such cyber-attacks, they add a huge burden to the CAs and OCSP servers. Indeed, CAs and/or OCSP servers need to timestamp and sign on a regular basis all the responses, for every certificate they have issued, resulting in a very high overhead. To mitigate this and provide a solution to the described cyber-attacks, we present CCSP : a new approach to provide timely information regarding the status of certificates, which capitalizes on a newly introduced notion called Signed Collections . In this article, we present in detail the notion of Signed Collections and the complete design, implementation, and evaluation of our approach. Performance evaluation shows that CCSP (i) reduces space requirements by more than an order of magnitude, (ii) lowers the number of signatures required by six orders of magnitude compared to OCSP-based methods, and (iii) adds only a few milliseconds of overhead in the overall user latency.

In recent years, Header Bidding (HB) has gained popularity among web publishers, challenging the status quo in the ad ecosystem. Contrary to the traditional waterfall standard, HB aims to give back to publishers control of their ad inventory, increase transparency, fairness and competition among advertisers, resulting in higher ad-slot prices. Although promising, little is known about how this ad protocol works: What are HB's possible implementations, who are the major players, and what is its network and UX overhead?