Inside the AI-powered assault on SaaS: why identity is the weakest link (www.techradar.com)

Recent findings from AppOmni's State of SaaS Security 2025 Report reveal a troubling trend where cyber attacks increasingly start with stolen identities rather than traditional malware. With 75% of organizations experiencing SaaS-related incidents in the past year, many attackers now leverage AI to impersonate legitimate users, allowing them to bypass security controls and operate within trusted environments unnoticed. This shift highlights identity as the new perimeter in cybersecurity, as compromised credentials—including passwords, API keys, and OAuth tokens—serve as gateways to sensitive data. The implications for the AI/ML community are significant, as threat actors are utilizing sophisticated AI techniques to enhance their attack strategies. This includes automating reconnaissance processes to gather invaluable contextual information about potential targets and analyzing vast datasets of stolen credentials to identify high-value accounts for subsequent attacks. Moreover, the mass production of synthetic identities using AI tools poses a new dimension of threat, complicating traditional security measures. To combat these emerging risks, organizations must rethink their security strategies by placing identity at the core of their defenses and leveraging AI capabilities to distinguish between genuine and synthetic users in real time.