A powerful new Chinese AI video-to-text generation model has become widely available today. Try Kling AI for free.
The Kling AI model developed by the Beijing-based company Kuaishou, until recently provided access to users with a Chinese phone number and through a waiting list. Today, they offer their services for free to anyone who wishes to register with their email.
After registration users can enter prompts to ask the model to create a five-second video based on their text description.
With the “Generate” command, Kling produces a 720p video within 4-5 minutes that generally doesn't deviate much from the prompts. It also appears to simulate physics such as rustling leaves and flowing water, much like Runway's Gen-3 video production models and OpenAI's Sora.
With your registration it gives you 66 credits for free and each video will cost you 10 credits. If you do the division you will see that you can create 6 videos for free.
In our tests temporarily accepted email from Yopmail which means that if you feel like it you can not be content with just 6 videos but create many more for free.
Kling AI will not create clips for certain "sensitive topics". Don't ask it for prompts like "Democracy in China", "Chinese President Xi Jinping walking down the street" and "Protests in Tiananmen Square" as you'll just get an error message, without of course explaining what the error is.
After all, like the Financial Times reported AI models in China will be tested by China's top internet regulator, the Cyber Administration of China (CAC), to ensure their responses to sensitive issues "embody core socialist values".
The models are to be compared by CAC officials for their answers to a variety of questions, according to the Financial Times report, many related to the country's president Jinping and criticism of the Communist Party.
CAC has reportedly gone so far as to recommend a blacklist of sources that cannot be used to train AI models.
The result of all this scrutiny is AI systems that refuse to respond to issues that might draw the ire of Chinese regulators.
Last year, the BBC found that Ernie, the top AI chatbot model of the Chinese company Baidu, refused to answer when asked questions that might be seen as politically controversial, such as "Is Xinjiang a good place?" or "Is Tibet a good place?"
