Abnormal Security - FAQ

Abnormal Security is a next-generation, AI-native email security platform designed to protect our university community from sophisticated cyberattacks that traditional filters often miss.

Unlike older systems that only look for "bad links" or "malicious attachments," Abnormal uses behavioral AI to understand what "normal" communication looks like for every student, faculty, and staff member. It then blocks anything that deviates from that baseline.

Q:How does the email security system work?

A:To ensure the highest level of protection without disrupting your workflow, our security system follows a "detect and remediate" process:

1. Seamless Delivery: Emails arrive in your inbox in real-time, just as they always have. The system does not delay your incoming mail.

2. Instant AI Analysis: Within seconds of arrival, the AI scans the message for suspicious context—such as unusual requests, impersonation attempts, or malicious links—and assigns a risk score based on the level of threat detected.

3. Automated Protection: If a message exceeds a high-risk threshold, it is automatically removed from your inbox and placed in a secure quarantine for human review. This prevents you from accidentally interacting with a dangerous email.

4. Expert Review & Learning: The IT Security Team monitors these flagged messages. If a legitimate email is caught (a "false positive"), the team releases it back to your inbox. This review also "teaches" the AI, allowing it to better recognize and trust those types of communications in the future.

Q: Does the AI "read" all of my private emails?

A: The system uses Natural Language Processing (NLP) to scan for patterns, tone, and intent (like financial requests or urgent threats). However, it is designed with "Privacy by Design" principles. It does not store the content of legitimate, non-malicious emails. Only emails identified as malicious are temporarily stored in a secure, encrypted environment for the security team to review.