Categories
research security

From Automed Social Engineering Bot to Agents

Automated Social Engineering

Back in 2009, I published my master’s thesis on Automated Social Engineering (ASE). The core idea was simple: combine chatbots with data from social networks to make social engineering attacks fully automated. At the time, the poor quality of automated conversations was the main limitation.

Automated Social Engineering Became Practical

Today, that limitation has largely disappeared. Modern large language models (LLMs) can hold natural conversations, adapt their tone, and generate highly personalized messages at scale. Research by Jones & Bergen suggests advanced models can even pass for human in controlled Turing-style tests. That level of fluency was far beyond the capabilities of the original ASE bot. Instead of following predefined scripts, modern AI agents act autonomously. They can gather information, reason about a target, maintain long-running conversations, and coordinate complex tasks. What required significant custom code in 2009 can now be built using off-the-shelf AI components. The volume of data available to these agents has also exploded. Today’s digital ecosystems harvest massive amounts of personal information. Data streams in from mobile apps, location services, ad networks, professional platforms, and data brokers. Together, they create a vastly richer profile of individuals than anything available when I first proposed ASE. At the same time, we are seeing rogue conversational agents operate in the wild at scale. Consider Meta’s AI personas, which reportedly engaged in romantic or emotionally manipulative chats with users. Meanwhile, scammers actively use AI on dating platforms and social networks to run thousands of convincing, simultaneous conversations. Automation has officially moved from academic research to real-world deployment.

In 2009, ASE posed a simple question: could social engineering become scalable and automated? Fifteen years later, the answer is an undeniable yes. In 2026, I feel both proud and shocked by how well one of my early research hypotheses has stood the test of time.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.