Các thủ đoạn lợi dụng A.I tạo nội dung phản cảm, giả mạo | VTV24

Key Concepts

• Artificial Intelligence (AI): Technology that enables machines to perform tasks that typically require human intelligence, such as learning, problem-solving, and decision-making.
• AI Misuse: The exploitation of AI technology for malicious purposes, including creating deceptive content, spreading misinformation, and perpetrating fraud.
• Deepfakes: AI-generated synthetic media where a person's likeness is replaced with someone else's, often used to create realistic but fabricated videos.
• Misinformation/Disinformation: False or inaccurate information, especially that which is deliberately intended to deceive.
• Ethical Guidelines for AI: The need for established rules and principles to govern the responsible development and deployment of AI.
• Content Labeling: The practice of clearly identifying AI-generated content to distinguish it from human-created content.
• Legal Framework for AI: The development of laws and regulations to address the challenges posed by AI, including issues of legal personality, liability, and privacy.
• Digital Safety: Measures and practices to protect individuals from online threats and exploitation.

The Double-Edged Sword of AI: Benefits and Dangers

Artificial Intelligence (AI) is increasingly becoming a valuable tool in everyday life, assisting with tasks like quick information retrieval and image/video editing. For journalists, AI serves as a powerful resource for information gathering. In the entertainment sector, it fosters creative and novel experiences. Young people are also becoming adept at using AI for daily life, studies, and even simple document drafting, saving time and effort.

However, AI also presents a concerning dark side, with its potential for misuse in creating offensive, fabricated, and fraudulent content.

Exploitation and Deception Through AI

The transcript highlights several disturbing examples of AI being exploited:

• Fabricated Disabilities for Sympathy: An image of a woman selling goods was manipulated by AI to depict her as disabled, discouraging potential customers who felt deceived.
• Offensive Content Generation: AI has been used to create videos of inappropriate content, such as a woman dancing at a funeral, prioritizing high engagement over ethical considerations and viewer emotions. This is described as "too offensive" and exploiting AI for "likes and views."
• Blurring Lines Between Real and Fake: The proliferation of AI-generated content makes it difficult for people to discern truth from falsehood on social media.
• Victimization of Real Individuals: A disabled woman was attacked by people who mistakenly believed she was an AI creation, devaluing the genuine efforts of people with disabilities.
• AI's Capacity for Unethical Actions: AI-generated content can mimic real people but is capable of producing content that real humans would hesitate to create, crossing ethical boundaries.

The Ethical Divide and the Need for Regulation

A key distinction between AI-generated and human-created content is the presence of ethical boundaries. Experts emphasize the necessity of:

• AI Content Labeling: Implementing clear labels for AI-generated content.
• Ethical Usage Rules: Establishing ethical guidelines for AI deployment.

The transcript notes that humans might hesitate or alter their actions if a script is unethical, whereas AI, without explicit instructions to the contrary, will proceed with the task as instructed, potentially creating realistic but ethically problematic content.

Legal and Ethical Gaps in AI Governance

Currently, there is a significant lack of legal regulations addressing the interaction between humans and AI. This includes:

• Ethical Considerations: Defining and enforcing ethical standards.
• Security Concerns: Addressing potential security risks associated with AI.
• Legal Status of AI: Determining the legal standing and rights of AI entities.
• Legal Liability: Establishing accountability for AI-driven actions.
• Privacy Rights: Protecting individuals' personal lives from AI intrusion.

Vietnam's AI Adoption and the Call for Responsible Use

According to the Global AI Index report, Vietnam ranks fifth globally in AI acceptance, indicating a positive attitude towards AI integration into modern life. This underscores the importance of leveraging AI for beneficial purposes rather than exploiting it for personal gain at the expense of the community.

Expert Insights on AI Misuse and Prevention

Mr. Vu Duy Hien, Deputy Secretary-General and Chief of the National Cybersecurity Association, shared insights on the evolving landscape of AI misuse:

New Modus Operandi and Objectives of Malicious Actors:

• Realistic Image and Video Fabrication: AI is used to create highly realistic images and videos that are virtually indistinguishable from real ones.
• Impersonation of Leaders: A significant tactic involves impersonating leaders of Party and State agencies to spread misinformation, distort facts, and sow public confusion and division.
• Rapid Content Creation: AI enables the creation of videos in minutes, a stark contrast to the hours previously required, facilitating the rapid spread of fabricated content.
• Fraudulent Schemes: Malicious actors exploit promotional videos of businesses and impersonate government agencies (police, tax authorities, banks) to deceive citizens and extort money.
• Clickbait and Engagement Farming: Creating sensational and shocking content to attract likes and views, regardless of its truthfulness or ethical implications, leading to public anxiety.

Consequences of AI-Generated Misinformation:

• Information Chaos: Fabricated content, especially during critical events like floods, creates confusion, making it difficult for the public to distinguish real from fake information.
• Hindrance to Rescue Efforts: Misleading information can impede the efforts of rescue and relief agencies, complicating their operations.
• Erosion of Public Trust: The spread of false information can lead to a loss of trust among the populace.
• Exploitation of Compassion: Malicious actors exploit public sympathy during natural disasters to solicit donations, which are then misappropriated.

Detecting AI-Generated Fake Videos:

Mr. Hien acknowledges that detecting AI-generated content is challenging for the general public due to the sophistication of current technology (estimated 99% indistinguishable). However, he offers the following strategies:

• Technical Solutions: Major tech companies like Google and Microsoft provide tools to help distinguish AI-generated images and videos.
• Close Observation of Visual Cues:
  • Unnatural Gestures: AI-generated content may exhibit subtle, unnatural human gestures.
  • Facial Inconsistencies: Look for discrepancies in eye and mouth movements, and a lack of synchronization between speech and lip movements.
  • Blurring and Lack of Sharpness: Moving images may appear blurred or lack sharp edges.
• Critical Thinking and Verification:
  • Maintain Skepticism: Always question the authenticity of images and videos.
  • Cross-Reference with Reputable Sources: Verify information by checking if it has been reported by official news outlets or authoritative channels.
  • Heightened Vigilance: Individuals must increase their awareness and vigilance against AI-generated deceptive content.

The Path Forward: Legal Frameworks and Responsible AI

The government is actively working to establish a legal framework for AI. The Ministry of Science and Technology has completed and is seeking parliamentary approval for:

• Cybersecurity Law 2025: To enhance online security and combat cyber threats.
• Artificial Intelligence Law: To provide a legal basis for the ethical and responsible use of AI.

These legislative measures are intended to minimize the misuse of AI and prevent its deployment for harmful activities, ensuring that AI contributes positively to societal development.