1. <td id="kfski"></td>

      <track id="kfski"></track>

      <track id="kfski"></track>

      <p id="kfski"></p>

      position: EnglishChannel > News > Article

      Text-to-Video Generator Sora a Mixed Blessing

      Source: Science and Technology | 2024-02-21 15:55:30 | Author: Tang Zhexiao

      OpenAI recently announced Sora artificial intelligence, which can transforms text into video of up to 1 minute. (PHOTO: VCG)

      OpenAI, the creator of ChatGPT and image generator DALL-E, launched a new artificial intelligence (AI) tool that enables users to create short videos from text prompts on February 15.

      Named "Sora," this AI-video tool can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions, OpenAI said.

      However, the San Francisco-based startup admitted that the new tool still has some limitations, such as possibly "mixing up left and right", according to AFP.

      The technology that supports Sora is an adaptation of DALL-E. It generates a video by starting off with noise and "gradually transforms it by removing the noise over many steps," the company explained. It recognizes objects and concepts listed in the written prompt and pulls them out of the noise, so to speak, until a coherent series of video frames emerge.

      The impact of Sora in shaping video generation and its implications for various industries has been seen through factors like enhanced text-to-video capabilities and exploration of novel applications.

      According to AFP, the French video game giant Ubisoft hailed the tool as a "quantum leap forward" with the potential to let players and development teams express their imaginations.

      "For professions like marketing or creative, multimodal models could be a game changer and could create significant cost savings for film and television makers, and may contribute to the proliferation of AI-generated content rather than using actors," Reece Hayden, senior analyst at a tech intelligence company ABI Research, told CBS MoneyWatch.

      Besides the praise by some AI researchers, concerns about security were also raised.

      "The video generation model is spurring excitement about advancing AI technology, along with growing concerns over how artificial deepfake videos worsen misinformation and disinformation during a pivotal election year worldwide," said New Scientist.

      Hany Farid, professor at the University of California, Berkeley, specializing in image analysis and digital forensics, said "text-to-video will continue to rapidly improve — moving us closer and closer to a time when it will be difficult to distinguish the fake from the real."

      The new video tool is not yet publicly available. OpenAI has restricted its use to "red teamers" and some visual artists, designers and filmmakers to test the product and deliver feedback before it is released more widely.

      Editor: 湯哲梟

      Top News

      • The exports of China's new energy vehicle (NEV) sector soared by 77.6 percent in 2023, reaching 1.2 million units, ranking first in the world for the ninth consecutive year.

      Joint Efforts to Build World's Largest Radio Telescope

      China was right there in the early conversations about what the future of radio astronomy might look like, and has played a critical role in creating the world's largest and most advanced radio telescope - the Square Kilometer Array (SKA).

      Report Details China’s High-quality Growth in Five Sectors

      A new report, titled China’s "Compounding Interest”: High-Quality Development and Outlook for 2035", was released at an international symposium in Beijing on March 31. Compiled by five think tanks in collaboration, from China, the U.S., Russia, Canada and India, the report illustrates the achievements of China’s high-quality development in five major sectors in the past few years with detailed data.

      JAPANESE国产高清在线观看_视频小说图区欧美国产_久99久爱精品视频_国产三级精品三级男人的天堂
      1. <td id="kfski"></td>

        <track id="kfski"></track>

        <track id="kfski"></track>

        <p id="kfski"></p>