Rising shadow AI usage and pressure to relax controls expose critical risks despite high enterprise confidence
India is emerging as a global leader in AI adoption confidence, yet a new report from Delinea highlights a widening gap between perception and reality when it comes to identity security governance. According to the study, “Uncovering the Hidden Risks of the AI Race,” 96.8% of Indian organizations believe their identity security posture is ready to support AI at scale. However, 57.6% simultaneously identify AI environments as their weakest area of governance the highest globally.
The research, based on insights from over 2,000 IT decision-makers across seven countries, underscores a growing “AI security confidence paradox” in India. While 51.6% of Indian respondents consider themselves “very prepared” for AI-driven transformation well above the global average only a fraction demonstrate real-time visibility into emerging risks.
“India is one of the most enthusiastic adopters of agentic AI in the world, but our research shows that Indian organisations are also accepting more identity risk to fuel that speed than almost any other country we surveyed,”
Anand (Jude) Kannabiran, Vice President, Asia at Delinea.
One of the most concerning findings is the rapid rise of shadow AI. Nearly 68% of Indian organizations reported encountering unauthorized AI tools or agents accessing enterprise systems over the past year, significantly higher than the global average of 53%. Alarmingly, only 27.6% said they can detect such activity in real time.
The report also reveals mounting internal pressure on security teams. Over 93% of Indian respondents admitted facing demands to relax privileged access controls to accelerate AI adoption the highest among all surveyed countries. This often results in compromised security practices, including granting standing access or bypassing controls altogether.
Further compounding the issue, 42% of organizations still rely on static, long-lived credentials for managing AI and non-human identities, increasing exposure to breaches. The business impact is already evident only 7% of respondents reported no negative consequences from identity-related friction, with delays in AI initiatives and rising operational costs among the most cited challenges.
The findings highlight a critical inflection point for Indian enterprises. As AI agents increasingly interact with sensitive systems and data, organizations must shift from confidence-driven adoption to governance-led innovation. Strengthening identity visibility, enforcing least-privilege access, and adopting real-time monitoring will be key to securing the next phase of AI growth.