Baidu's pioneering technology work recognized by World Internet Conference

A hyper-realistic digital human technology developed by Chinese tech giant Baidu Inc won the World Internet Conference Award for Pioneering Science and Technology in Wuzhen, Zhejiang province on Thursday.

As a result, Baidu has become the only artificial intelligence company to win at the awards for three years running. Initiated by the World Internet Conference, the awards aim to recognize pioneering achievements in internet technology globally and promote international exchanges and cooperation in the internet domain.

AI foundation models are rapidly evolving from language to multimodal capabilities, with "digital humans" – 3D computer generated representations of human beings which can act like real people in virtual environments such as broadcasts – emerging as an innovative application that integrates large language models with multimodal technology.

Baidu won this year's award with a script-driven multimodal collaboration hyper-realistic digital human technology, featuring innovative techniques such as script generation integrated with multi-modal planning and deep reasoning, real-time interaction through dynamic decision-making, text-controlled speech synthesis and high-consistency ultra-realistic long video generation of digital humans.

This breakthrough technology addresses key challenges such as real-time multimodal coordination and comprehensive dynamic interactions, achieving the consistency of voice, appearance and speech, the company said.

Baidu says it aims to deliver high-quality content and more natural interaction for digital human live broadcasting, characterized by sophisticated expressiveness, strong emotional impact, free interaction among people, objects and environments, and professional descriptions.

To date, more than 100,000 digital humans have been produced using this technology, covering industries such as e-commerce, education and law. It has helped reduce the cost of starting a live-broadcast by 80 percent and increased the conversion rate by 31 percent.

This technology plays a key role in driving innovation in the live broadcasting industry and lays a solid foundation for the widespread application of digital human technology.