During this year's National Day holiday, some netizens utilized AI technology to clone Lei Jun's voice, using it to criticize uncivilized behavior or express dissatisfaction with others. Some joked that they had been "cursed at by AI Lei Jun for the entire holiday." In response, Lei Jun released a video today saying that it has indeed troubled and discomforted him. He expressed hope that people would stop this practice, deeming it inappropriate, though he still acknowledges AI as a valuable and practical technology.
(https://www.bilibili.com/video/BV1d6S4YPE7c?share_source=copy_web)
Recently, I participated in a test focused on AI videos, and the results shocked me. As a self-media creator who specializes in AI, I only scored 60 points while attempting to discern the authenticity of ten videos. I found it quite challenging. Interestingly, my friends had varying scores, with the highest being just 90 points. This experience made me acutely aware of how complex AI video production technology has become, making it hard to distinguish between real and fake videos.
With the rise of AI technology, fake video scams have proliferated. For example, news reported on an 82-year-old retired man who was duped out of \$690,000 of his life savings by a forex company supposedly endorsed by "AI Elon Musk." Such incidents are increasingly common online.
There was even a case where a Bilibili content creator used AI lip-sync technology and successfully tricked his mother, resulting in a loss of 20,000 yuan.
In July, a fabricated video of Biden circulated on the social platform X, which was later exposed as counterfeit AI. While the revelation elicited laughter and tears, the underlying security risks are serious.
Today, the barrier to creating AI videos is almost nonexistent. Last year, skepticism surrounded AI face-swapping and lip-sync technologies, but within just a year, these tools have become simple and efficient. Ordinary users can easily generate realistic AI videos by uploading a simple clip to an app and adding voiceovers. This trend is especially prominent on short video platforms like Douyin, where various entertaining yet unreliable AI content is prevalent.
As we discuss AI videos, it's also worth mentioning that XXAI is gearing up for an epic upgrade next week! This update will enhance existing features and introduce popular AI models like GPT-o1, perplexity, Grok, and Gmini, allowing users to experience a variety of AI capabilities for just \$9.9 a month, making advanced technology more accessible.
In the face of the rampant misuse of AI for creating fake videos, learning how to identify them has become increasingly important. I believe there are two simple and practical methods worth sharing:
Imagine an actor performing a puppet show. In a recording, if you closely observe the puppet's lips, you might notice blurriness around them, along with subtle tremors during voice synchronization. Currently, AI technology struggles to achieve perfect lip synchronization, especially regarding the appearance of teeth, which may appear duplicated or unnatural. By paying attention to these details, you can better discern authenticity.
This method may not be very helpful for content creators who frequently update their videos, but it can be effective for many 2D digital avatars. When AI avatars are involved in conversations, they often resort to repeating a few fixed movements and may exhibit inappropriate gestures or wandering gazes. Noticing these subtle cues can help identify unnatural AI-generated content.
As AI video technology matures, the issue of fake videos will only escalate on social media. Current regulatory frameworks have not kept pace with technological advancements, providing opportunities for wrongdoing. Enhancing the public's awareness and understanding of AI technology is crucial in combating misleading information, and each individual bears the responsibility of improving their discernment skills. In this fast-paced changing era, we must stay vigilant to recognize the truth rather than accept information at face value. Protecting ourselves and others while embracing technology is true wisdom.