
Ben Risher
author title Offensive Security Engineer LeadI am an Offensive Security Engineer and lead ExploitAI, a small team of multi-disciplined engineers operating at the intersection of Security, Artificial Intelligence, and Machine Learning. My team and I use our combined skills to assess Salesforce's models, products, and features to identify and help mitigate technical, reputational, and legal risk.
Before coming to Salesforce, I was engaged in zeroday research and reverse engineering with focuses on fuzzing, symbolic execution, and llvm/compiler internals. Additionally, I've spent time red teaming, both as an operator and a tool developer. I also was an instructor, teaching a six month long cybersecurity course for United States Air Force personnel. Alongside day-to-day work, I enjoy giving back to the security community via open source development. I've contributed code to many open source projects such as the metasploit framework, exploit-db, and LibAFL. I'm also the author of feroxbuster, a popular offensive security tool designed to unearth hidden/unlink content on websites.
I'm either spending time with my family or exercising.


Salesforce's trusted AI architecture for red teaming leverages automation to scale ethical AI testing, utilizing a tool called fuzzai to simulate diverse adversarial scenarios and enhance model robustness. By automating adversarial prompt generation and response validation, fuzzai helps secure AI interactions while reducing human exposure to harmful content.

This blog details the process behind Salesforce’s first-ever AI Red Teaming Hackathon to build safety into our xGen family of AI Models.