Abnormal AI Launches Attune 1.0 to Combat AI-Driven Cyberattacks

Abnormal AI has introduced Attune 1.0, a behavioral foundation model designed to detect and block AI-driven cyberattacks that mimic normal workplace communication. This model, which accounts for 85% of detections on the Abnormal Behavior Platform, was trained on over one billion behavioral signals. The rise of generative AI has led to increasingly sophisticated attacks that personalize messages to target organizations. Attune 1.0 employs a multimodal architecture, integrating identity, behavior, and content signals to enhance detection capabilities. The system has reportedly increased detection of attack campaigns by 150,000 per week and improved precision by 50%. Recent updates to the platform include features aimed at enhancing visibility and administrator control in cloud email security. A notable success was the identification of a Microsoft Teams OAuth phishing campaign before it was publicly documented.

Key Points: • Attune 1.0 detects AI-driven cyberattacks with 85% detection rate across the platform. • The model was trained on over one billion behavioral signals to enhance detection capabilities. • Abnormal AI reported a 50% increase in detection precision, reducing false positives.