Original title: AI-Enhanced Email Security: A Novel Pipeline for Phishing Campaign Detection and Profiling
Tarini Saka, Kami E. Vaniea, Nadin Kökciyan · September 2026
Phishing attacks are getting harder to stop. Attackers now send thousands of coordinated emails using sneaky tricks, clever social engineering, and even artificial intelligence to sneak past filters. IT and security teams spend huge amounts of time reading, sorting, and stopping this flood of evolving threats. To help these workers instead of replacing them, researchers built a new artificial intelligence system that works in three steps. First, it uses fine-tuned language models to read phishing emails and pull out key contextual details. Next, it uses community detection algorithms to group similar emails into related attack campaigns. Finally, it creates attack summaries that explain the full campaign. By organizing related emails and writing these clear profiles, the system allows security teams to find and stop threats faster. As a result, personnel can block incoming malicious emails and quickly answer affected users.
Imagine sorting through a pile of letters to find ones with the same sender. Normally you read each letter individually, which takes a long time. Now suppose the letters are all slightly different but come from the same scammer. In this paper, the system automatically reads the letters, groups them by similarity, and writes a single summary for each group, so you can respond to the whole batch at once.
Security teams are overwhelmed by the volume and sophistication of phishing attacks. This pipeline helps them quickly identify coordinated campaigns and understand the attack's scope, reducing response time from hours to minutes. It allows defenders to block entire campaign waves instead of handling each email individually.
The abstract does not present experimental results or quantitative performance metrics. It does not compare the proposed pipeline against existing methods or demonstrate actual impact on response times. The approach is described conceptually, and no real-world deployment or dataset is mentioned. The claim of enhanced efficiency is based on the design, not on measured outcomes.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
AI-Enhanced Email Security: A Novel Pipeline for Phishing Campaign Detection and Profiling — Tarini Saka, Kami E. Vaniea, Nadin Kökciyan — ACM Transactions on Internet Technology — doi:10.1145/3848625
Abstract licence: CC BY
Explanation generated by Synthesis from the abstract — it is not the original paper.