Concerns about impersonation by AI
The explosion of generative AI (GenAI) has pushed the ability to create fake news and deepfaked news to a new level, making the boundary between reality and virtuality extremely fragile. AI (artificial intelligence) does not stop at simply editing images; this technology can fake voice from just a few seconds of sample data and create fraudulent videos with facial expressions, lip movements that fully match sound.
According to forecasts in 2026, up to 90% of online content can be created by AI. The World Economic Forum's Global Risks Report 2026 also ranks misinformation and distorted information as the second biggest threat in the next two years of the AI era.
Economic losses from fake news and fraud caused by AI in Vietnam are no longer theoretical figures. According to statistics from the Ministry of Public Security and the National Cyber Security Association, in the first 11 months of 2025, total damage from online fraud in Vietnam is estimated at 6,000 billion VND.
According to Mr. Jay Reddy - Head of Growth Department of cybersecurity company, ManageEngine technology solution in Vietnam, AI-based identity forgery technology is creating a crisis of trust for traditional biometric authentication systems. Previously, biometric factors such as facial features or voice were considered the most immutable and difficult-to-forge "keys".
However, the emergence of GenAI has turned these unique features into data that can be manipulated or completely copied. Electronic identification systems (eKYC) are facing the risk of being compromised when cybercriminals use sophisticated tools to bypass the life-detector layer. This makes it almost impossible for computers to distinguish between a real person present and a video recreated by AI with a latency of almost zero.
Solution
To optimize the power of AI in network security, ManageEngine experts have proposed 4 solutions for digital identity management in Vietnam - an important problem for business network security.
Establishing a standard framework for safe digital identification: Vietnam needs to quickly complete technical standards to build a foundational identification structure, providing continuous visibility of who has access rights and what they can do with those rights.
Promoting intelligence sharing models about threats: Businesses and government agencies should form a network to share information about the latest AI attack models in real time. Collective strength will allow us to identify and prevent cross-border AI fraud campaigns faster than the speed at which criminals change malware.
Manage the use of AI from within: AI security is not just a defense against external threats. Organizations must establish an internal governance framework to determine how to apply, monitor and control AI tools.
Investing in "digital people": The most important recommendation is to improve information evaluation capabilities for both experts and end users. Although technology can prevent 99% of attacks, the remaining 1% depends on human vigilance. An AI agent is only reliable when there is close supervision from humans.
