Cybersecurity firms SecureBank and Secure Innovation announce technical advancements

SecureBank has advanced its proprietary Cyber AI research to its third phase, focusing on distinguishing between productive re-verification and repetitive cycles that yield no new information. Through its ‘Information Gain Control Phase 1’ initiative, the company successfully connected an ‘attack reasoning chain’ within a CyberGym task. This process involved analyzing root causes, examining code differentials between vulnerable and patched versions, and constructing hypotheses to successfully reproduce a heap-buffer-overflow vulnerability.

Additionally, SecureBank conducted post-hoc analysis of AI execution logs from four real-world cases. The research aims to classify AI behavior into categories such as ‘PROGRESS’, ‘REVALIDATION’, ‘LOW_GAIN’, ‘STAGNATION’, and ‘POLICY_VIOLATION’ to determine if iterative testing is generating new evidence or merely repeating unproductive actions.

Separately, Secure Innovation has released an archive video of its seminar regarding vulnerability assessment lifecycles. The training covers selecting assessment targets, planning, utilizing results, and establishing continuous improvement cycles to protect web sites, business systems, and networks from cyberattacks.