Rapid AI Adoption Outpaces Enterprise Security Policies, Survey Finds

Nearly half of organisations lack a formal AI strategy even as more than 60 per cent of workers use AI tools daily, according to a new Okta survey that highlights a growing governance gap in enterprise AI adoption. The survey of nearly 800 executives and knowledge workers found that 58 per cent of organisations experienced an AI-related security incident in the past year, with employees granting AI tools broad access to email, files, calendars, and workplace collaboration tools without clear guardrails.

The data paints a picture of an adoption curve running far ahead of risk management. While 47 per cent of executives said their company lacks a formal AI strategy — down from 90 per cent in 2025 but still a critical gap — the disconnect between leadership and frontline workers is even starker. Thirty-five per cent of executives said AI policies are unclear, compared to 57 per cent of knowledge workers, a 22-point gap that suggests executives are significantly underestimating how ambiguous their AI rules actually are on the ground. Autonomous AI agents are already in widespread use across 58 per cent of organisations surveyed, making them the most common AI tool type ahead of chatbots and large language models.

The risk profile is sobering: 51 per cent of employees have shared work emails or internal messages with AI tools, 29 per cent have uploaded confidential company documents, and 16 per cent have entered login credentials. A separate LayerX Security survey found that while most users stick to a single AI assistant, the average enterprise employee now interacts with 2.24 different AI applications — multiplying the potential attack surface. ChatGPT commands 36 per cent of users and 55 per cent of conversations, followed by Microsoft Copilot at 29 per cent of users.

The findings come as the AI model risk management market expands to meet demand, projected to grow from US$7.17 billion in 2025 to US$15 billion by 2030, driven by mandatory AI compliance regulations and increasing demand for transparent, auditable AI systems. For Singapore enterprises, where AI adoption has been aggressively pushed through national initiatives like the National AI Strategy 2.0 and the Smart Nation programme, the governance gap carries particular weight — rapid deployment without corresponding guardrails can erode the very trust that Singapore's AI strategy depends on.

Why it matters for Singapore: Singapore has one of the highest rates of enterprise AI adoption in Asia, driven by government policy, generous grants, and a business ecosystem that rewards digital transformation. But the Okta survey is a reminder that adoption velocity without governance maturity creates real risk. For Singapore-based enterprises — many of which operate in regulated sectors like finance, healthcare, and government-linked services — the 58 per cent incident rate is a call to action. The city-state's AI ambitions will only hold if trust in enterprise AI deployment keeps pace with the tools themselves. The AI model risk management market's projected growth to US$15 billion suggests this isn't just a compliance issue — it's a business opportunity Singapore's burgeoning AI ecosystem can capture.