Meta, the parent company of Facebook, has recently entered the fray of AI video generation with the launch of Movie Gen. This new tool allows users to create videos with accompanying audio using a simple text prompt. Additionally, Movie Gen has the capability to edit videos based on user instructions. Meta claims that Movie Gen outperforms similar models in the industry when evaluated by humans.
The press release from Meta touts Movie Gen as the "most advanced and immersive storytelling suite of models," offering features such as video generation, audio generation, personalized video generation, and video editing. The models were trained using publicly available data and licensed data, according to the company.
One example provided by Meta shows a video of a baby hippo swimming underwater, while another depicts a koala bear surfing. Users can upload a picture of themselves to be incorporated into personalized videos, and Movie Gen can edit videos based on text instructions.
Meta's entry into the video generation space comes after competitors like OpenAI and Google have already established themselves in the market. OpenAI launched its video generator, Sora, in February, while Google followed suit with Veo in May.
Despite being a latecomer to the video generation game, Meta has made strides in the AI arms race. Recent iterations of Meta's Llama model have been well-received, with some viewing them as superior to OpenAI's offerings. Meta claims that its new models outperform competitors in human comparisons, with users preferring Movie Gen over OpenAI's Sora.
While Meta did not provide a direct comparison with Google's Veo, the company believes that its video-to-audio generation capabilities may surpass those of Google. However, Meta, OpenAI, and Google have not responded to requests for comment on their respective AI video generation tools.
Overall, Meta's Movie Gen represents the company's latest effort to establish itself as a leader in AI technology, despite entering the race later than some of its competitors.