Combating fake news in the AI era

Efforts to build a “digital shield”

The Politburo’s Resolution No.57-NQ/TW, dated December 22, 2024, on breakthroughs in science, technology, innovation and national digital transformation has brought about significant changes in awareness and action throughout the political system. It has driven tangible progress in improving institutions, developing digital infrastructure, reforming public services and advancing the digital economy, gradually shaping the foundations of a digital nation. However, as digital transformation accelerates, Viet Nam is also confronting major risks and challenges in ensuring cybersecurity.

In reality, AI, big data and deepfake technologies are, in themselves, achievements of science and technology. The issue lies in how and for what purposes they are used. When exploited for impersonation and distortion, these technologies become tools that significantly amplify risks in cyberspace.

According to Le Thanh Tung, Head of the Department of Communications and Knowledge Dissemination under the Viet Nam Union of Science and Technology Associations, producing and spreading fabricated content once required considerable time. Today, AI can generate large volumes of text, images, audio and video within a very short period. Deepfake technology can create images and voices that closely resemble real people, producing what may be described as “fake evidence that appears authentic”, making it easy for audiences to believe such content before they have time to verify it.

In particular, big data and content-distribution algorithms can enable information to be targeted at groups with different psychological profiles, interests or circumstances. False information that strongly appeals to emotions can be shared rapidly and repeatedly, creating the impression of a “truth widely accepted by many people”. New technologies therefore require new approaches: earlier detection, faster verification and more timely responses. At the same time, accurate, well-founded and convincing information must be provided proactively. Information gaps must not be allowed to become fertile ground for fake news and distorted narratives.

Against the backdrop of rapid technological development, it is therefore necessary to build a “digital shield” and apply AI to automatically scan for and issue warnings about harmful content. An effective “digital shield” should have at least three capabilities.

First is early warning: identifying unusual topics, sources of dissemination, the speed at which content is spreading, and signs of impersonation or coordinated manipulation. Second is rapid verification: connecting competent authorities, databases, experts and the press to verify information. Third is proactive communication: once a conclusion has been reached, accurate, accessible and multimedia content must be promptly produced and delivered to the appropriate audiences.

In this way, AI can do more than simply “detect harmful content”. It can also help analyse information needs, identify gaps in public understanding, suggest effective presentation methods, convert content into different formats, and distribute appropriate positive information.

However, according to Le Thanh Tung, all content must ultimately be verified by humans, who must also take responsibility for it; algorithms cannot be left to make such decisions unchecked. Resolution No.57-NQ/TW calls for mastery of technology and the development of digital infrastructure, data and innovation capabilities. Meanwhile, the Party Central Committee’s Secretariat’s Directive No.57-CT/TW, dated December 31, 2025, further calls for strengthened cybersecurity, information security and data security throughout the political system.

This demonstrates that a “digital shield” must form part of Viet Nam’s overall national digital capacity rather than being treated simply as a standalone software solution.

Further clarifying legal responsibility for AI-generated content

Viet Nam has made significant progress in strengthening the legal framework governing the digital environment. The Data Law took effect on July 1, 2025; the Law on Personal Data Protection came into force on January 1, 2026; and the 2025 Law on Cybersecurity took effect on July 1, 2026. These laws provide an important foundation for managing technological risks, protecting data, defining the responsibilities of relevant parties, and addressing violations in the digital environment.

The emergence of technologies such as deepfakes shows that the law must keep pace with technological advances, supported by enforcement mechanisms that are sufficiently swift, clear and deterrent. The greatest challenge, however, remains the fact that technology is evolving faster than the policy-making cycle. The first breakthrough must therefore be to establish an adaptable legal mechanism that clearly defines principles, responsibilities and risk thresholds, while providing mechanisms for rapid technical updates, controlled experimentation and inter-agency coordination.

According to Dr Nghiem Vu Khai, former Vice President of the Viet Nam Union of Science and Technology Associations, people are confronted every day with cross-border scams, impersonation of state agencies, fraudulent electronic contracts, financial investment and cryptocurrency schemes, AI-enabled impersonation, and personal data theft. In many cases, people do not intentionally violate the law but lack timely access to authoritative legal advice.

Priority should therefore be given to investing in a national AI-powered digital legal assistant system to help all organisations and individuals search for legal information, obtain answers to legal questions, receive warnings about legal risks, and access authoritative legal information quickly and accurately. At the same time, a national AI system should be developed, managed and operated to support the dissemination of and education about the law, ensuring that legal data are fully updated, information security and personal data protection are guaranteed, and the quality of AI-generated content is properly controlled.

Sharing this view, Le Thanh Tung said that efforts to combat fake news and safeguard digital sovereignty require further clarification of legal responsibility for AI-generated content. This includes the responsibilities of model providers, distribution platforms, deployers, content creators and disseminators, particularly when fabricated content affects social order and safety, infringes upon honour and dignity, or causes damage to organisations or individuals.

The continued improvement of the legal framework must safeguard the legitimate rights and interests of citizens and businesses, the right of access to information, and space for innovation. This is also in line with the spirit of Resolution No.57-NQ/TW, which calls for greater autonomy and self-reliance in science, technology and digital transformation.