IBM India Exec Highlights AI Security: Identity as the New Perimeter

The integration of artificial intelligence into cybersecurity has fundamentally altered how organizations must defend their networks. According to Gaurav Agarwal, Vice-President of Technology for IBM India and South Asia, traditional perimeter-based security is no longer sufficient as cloud services, remote work, and autonomous AI agents become standard. Instead, security frameworks must now center on identity verification, as attackers increasingly use valid login credentials to appear as legitimate insiders.

AI agents, which can execute tasks across multiple applications simultaneously, present unique risks. Without clear governance and strict adherence to the principle of 'least privilege,' these agents can be manipulated via 'rogue prompts' to expose sensitive data or facilitate unauthorized access. Agarwal emphasizes that Indian organizations must treat AI agents with the same security rigor as human employees, incorporating continuous monitoring, input validation, and detailed audit logging to contain potential vulnerabilities.

To counter attacks that strike with machine-like speed and volume, organizations must adopt a proactive, AI-based defense. By scanning networks for weak points before attackers do, companies can preemptively secure their infrastructure. Furthermore, Agarwal stresses the importance of keeping humans in the loop for strategic business decisions, noting that AI should be used to influence choices rather than override human oversight in critical processes, particularly as cybersecurity continues to evolve in the Indian and global landscape.