Confident execs face risk as most staff could be fooled by deepfakes

Nearly one-third of workers would use unapproved agentic AI if official tools are slow. 

Most (87%) employees said they could be fooled by a sophisticated deepfake scam posing as an internal stakeholder or executive, even as business leaders remain confident that staff can identify impersonation attempts.

A KnowBe4 report revealed that the figure is higher than the global average of 64%. It also showed that 93% of workers said deepfake content has become so realistic that it is harder to know what to trust.

Despite this, 88% of organisation leaders said employees can identify impersonation messages through internal tools. Another 74% said workers can identify deepfake voice and video content.

However, more than a third (37%) of employees were not confident in their ability to identify deepfake audio or video impersonations.

"Cybersecurity has entered a volatile phase where organisations are trying to secure a hybrid human and artificial intelligence (AI) workforce that’s changing more quickly than security leaders can keep up," said Kawin Boonyapredee, CISO Advisor at KnowBe4 APJ. 

The report showed that about 40% of surveyed Singapore organisations deploy autonomous AI agents, whilst 44% of leaders said AI use within their group is either unapproved or lacks formal governance.

Meanwhile, over half (56%) of the country’s cybersecurity decision-makers said unapproved external software and rogue AI applications had weakened or compromised their security posture over the past 12 months.

Approximately 56% of workers use unapproved tools, whilst nearly one-third (32%) would turn to unapproved agentic AI tools if official corporate tools are restricted or too slow. 

"Attackers are moving at machine speed, using attacks such as deepfakes to target employees and prompt injections to hijack AI agents. Leaving nearly half of your corporate AI usage ungoverned is a massive open invitation to threat actors,” Boonyapredee added.

Only 76% of security leaders feel prepared for unexpected or emerging AI-driven threats over the next year, below the global average of 92%.