Singapore Workers Fear Deepfakes as KnowBe4 Finds 44% of AI Use Ungoverned

Nearly every organisation in Singapore has suffered a human-related cybersecurity incident in the past year, according to a new study from security awareness firm KnowBe4. The survey of 200 Singapore employees and 50 security decision-makers reveals two intertwined problems: workers cannot reliably detect deepfakes, and nearly half of all corporate AI tool usage is operating without governance or approval.

The study, part of a global survey of 4,000 professionals across organisations with 250 or more employees, found that 93% of Singapore employees say deepfake-generated voice and video are now so realistic it is difficult to know what to trust—significantly higher than the global average of 86%. Even more striking, 87% of Singapore workers believe they could be tricked by a convincing impersonation scam at work, compared with 64% globally. These numbers place Singapore among the most exposed workforces to AI-generated deception, where internal messaging platforms and collaboration tools have become common attack vectors.

Beyond the deepfake trust crisis, the study highlights an alarming governance gap. While 99% of organisations in Singapore use AI in their workflows, a full 44% of that usage is unapproved or ungoverned. Forty percent of cybersecurity leaders report that AI agents already take fully autonomous actions inside workflows without direct human intervention. Meanwhile, 32% of employees admit to sourcing their own agentic AI tools when approved options are unavailable or too restrictive. More than half of cybersecurity leaders said unsanctioned software and AI applications had affected their security posture in the past 12 months.

Perhaps the most concerning finding is the disconnect between leadership confidence and employee behaviour. While 88% of leaders expressed confidence in their teams' ability to identify impersonation messages, 67% of employees say time pressure and distractions cause them to knowingly skip security procedures. Even more troubling, 37% of employees admit they sometimes choose not to report a security mistake because they feel embarrassed—despite 88% of organisations believing their employees feel safe reporting errors.

Why it matters for Singapore: As the city-state accelerates enterprise AI adoption through initiatives like the National AI Strategy and Smart Nation programmes, this study exposes a critical vulnerability: the human layer. Singapore's AI push has focused heavily on infrastructure, talent, and regulation, but the KnowBe4 data suggests that worker confidence in detecting AI-generated threats is lagging dangerously behind adoption. With Singapore positioning itself as a trusted AI hub, closing the gap between AI deployment and AI security literacy is not optional—it is foundational to maintaining that trust.