(Bloomberg) – Facebook parent company Metaplatforms Inc. is debuting a new artificial intelligence tool that can generate or edit videos based on simple text prompts, putting it ahead of OpenAI and Google in the race to develop the world’s most cutting-edge AI technology. competition with its rivals is intensifying.
Most Read Articles on Bloomberg
Movie Gen, a product of Meta, can create new videos up to 16 seconds long based on text prompts. You can also use these prompts to generate or edit audio for existing videos, or use photos to create customized videos featuring real people.
Movie Gen is only available to some internal employees and a small number of external partners, including some filmmakers, but the social media company, which also owns Instagram, WhatsApp, and Messenger, expects it to be available to existing We plan to incorporate this product into our app.
Executives are still debating the best way to implement this integration, but Connor Hayes, vice president at Meta, which specializes in generative AI products, said MovieGen’s goal is to encourage more people to post videos. The idea is to encourage creation and editing. It’s “fun to use, useful for creators, and good for overall app engagement, but we don’t have any concrete product plans for what that will look like at this point,” he said. Ta.
The social networking leader is one of many big tech companies pushing AI models focused on video generation, but not on models that return AI-generated text. more complex and expensive to build. Microsoft Corp.-backed OpenAI has its own video generation tool called Sora, which debuted earlier this year and can create videos up to a minute long, but the technology is not yet publicly available. DeepMind, a subsidiary of Alphabet Inc.’s Google, also has a video generation tool called Veo, which was announced earlier this year.
Meta is waiting to roll out Movie Gen for several reasons, including the efficiency of the technology. Currently, it takes “several tens of minutes” to generate a video based on a text prompt, Hayes said, which is too long for average consumers who are likely to use this on their phones.
But Hayes said Meta is also “sorting out a number of very important issues around safety and liability,” including videos that users can watch that feature other people without their consent. It said it also includes how to handle personalized videos to prevent the creation of offensive videos. “This is probably the most important issue to resolve before personalization capabilities are widely available to people.” It has been used in the past to create fakes. Meta executives said they are considering ways to “watermark” the works of Biden and pop star Taylor Swift so they can be identified as AI-generated.
story continues
Meta has made advancements in AI a key company-wide priority, with CEO Mark Zuckerberg repeatedly speaking about AI as a driver of user and revenue growth. In the short term, Zuckerberg credited AI with helping the company improve its content algorithms by showing people more relevant posts and ads. Ultimately, Zuckerberg believes AI will play an even bigger role in powering his company’s apps and future wearable products, such as the smart glasses being developed by Meta. He said there was.
Most Read Articles on Bloomberg Businessweek
©2024 Bloomberg LP
