Overview
Severity: CRITICAL | Affected: Adept AI Labs | Category: breach
Adept AI Labs, a leading AI research firm, confirmed a significant data breach that occurred in late May. Attackers gained unauthorized access to their internal network, exfiltrating terabytes of sensitive data. The compromised data includes proprietary source code, unreleased model weights for their next-generation 'Synergy-5' large language model, and a substantial portion of their curated training datasets. The company stated that customer PII was not affected, but the intellectual property loss is substantial. The breach was traced back to a compromised developer credential obtained via a sophisticated social engineering attack. This incident highlights the increasing targeting of AI companies for their valuable IP and the critical need for robust internal security controls beyond standard perimeter defenses.