Urgent need to control AI content

Criteria need to be issued to limit the content users are allowed to create with AI, forcing clear labeling

In just the past few years, artificial intelligence (AI) has developed and continuously improved at a dizzying speed, becoming a deep support tool in many fields. Veo3 - a video creation technology using AI recently developed by Google DeepMind - is capable of creating lifelike footage. Besides the utilities supporting users, this tool also poses the challenge of the risk of being exploited by bad guys for fraud.

Difficult to distinguish between real and fake

According to the technology world, Veo3 has "caused a stir" globally because of its AI video creation feature with a level of realism that surpasses any previous tool. Unlike OpenAI's Sora - which only stands out in terms of images but is limited in terms of sound, Veo3 perfectly synchronizes images, environmental sounds and dialogue. From footsteps, wind blowing to human voices and surrounding effects, all are smoothly combined, reaching a level of lifelikeness, making it difficult for users to distinguish between real videos or those created by AI.

In Vietnam, Veo3 quickly created a fever on social networks such as TikTok, Facebook and YouTube Shorts. A series of videos with the hashtag #veo3 were shared at a dizzying speed, attracting millions of views and hundreds of thousands of interactions. Many clips created by AI surprised the online community because of their unbelievable level of authenticity. For example, an 8-second video with a simple dialogue reached 5 million views and more than 308,000 likes. Another video recording an elderly woman frying eggs, created entirely by AI, also reached more than 2.6 million views and more than 104,000 likes.

Mr. Tran Vinh Khang - an office worker in District 1, Ho Chi Minh City - shared that he could hardly distinguish between real and fake because the authenticity was too high. "I saw a video of a fish seller talking to customers at the market and thought it was real because the image was so sharp. The way he waved his hand, raised his eyebrows, the sunlight shining on his face, everything was so natural that it was unbelievable that it was created by AI" - Mr. Khang said.

Images in the film created by Veo3, users find it difficult to distinguish between real and fake

Although she admitted to being impressed, Ms. Nguyen Mai Anh (living in Ho Chi Minh City) expressed concerns about information security. "The video created by Veo3 is so realistic that it makes me feel insecure. If bad guys use this technology to create videos to slander, defame or forge evidence, the consequences will be very serious" - Ms. Anh expressed.

Many Vietnamese users are particularly interested in how to access and use Veo3. However, Veo3 has only been officially deployed by Google in the US market, through the Google AI Ultra subscription package with a fee of 249.99 USD/month. Faced with this limitation, the Vietnamese online community has found a way to "overcome the barrier" to experience it. Many people share tips on using VPN to change IP to the US, thereby registering a Google account in a supported region to activate and use Veo3 as a local user. According to security experts, this is also a loophole that bad guys can exploit to do shady things.

Legal framework needed for AI

Mr. Le Hong Duc, founder of OneAds Digital Company Limited, said that AI-generated products such as videos, images or voices are approaching an astonishing level of realism, and can even be directly put into production and commercialization. This explosion brings clear benefits in terms of cost and time, but can be exploited to spread false information or cause public confusion. In the context of increasingly complex AI models, content control becomes urgent, especially when technology has far outpaced the speed of updating current laws.

Therefore, Mr. Duc proposed that it is urgent to develop a specialized law for AI, which clearly defines the responsibilities of those who create and share AI content. Acts of spreading false and harmful content must be strictly sanctioned and handled. At the same time, authorities need to issue criteria to limit the content that users are allowed to create with AI and require clear labels. If not implemented correctly, users must bear legal responsibility and compensate for any damage caused. "Issuing a separate legal framework for AI is extremely urgent to control this rapidly developing technology, contributing to maintaining a healthy digital environment, preventing consequences such as fraud, ethical violations or acts of insulting others" - Mr. Duc emphasized.

In addition to legal concerns, an information security expert shared that the outstanding development of Veo3 also raises concerns among content creators, especially in fields such as film, television and animation, which are at risk of being gradually replaced by AI technology, posing a major challenge to protecting careers and copyrights in the creative industry. "It is necessary to soon supplement regulations on intellectual property rights for AI-generated content, because AI is now able to produce highly realistic products. Otherwise, creators will face the risk of losing their jobs due to the wave of virtual works flooding the market," the expert said.

Beware of fake services

Since Veo3 exploded, on social media platforms, many accounts claiming to be "AI experts" have also started sharing demo videos, instructions on how to access Veo3 or selling accounts and support services for prices ranging from 200,000 to 300,000 VND. Some users said they had successfully accessed and tried it, but not everyone could experience it. Mr. Le Hong Duc warned users to be careful with fake or unofficial services, and recommended waiting for Google to officially deploy Veo3 in the Vietnamese market to avoid financial fraud, data theft or control of Google accounts.

Source: https://nld.com.vn/cap-bach-kiem-soat-noi-dung-ai-196250607190357821.htm