At this year’s Jiangsu TV New Year gala, singer Zhou Shen performed a duet across time with a digital recreation of Teresa Teng, a moment that spread widely online and showed what digital humans can do. At a Baidu tech salon on the theme, Li Shiyan, head of Baidu Intelligent Cloud’s AI human-computer interaction lab, said the platform’s goal over the next two years is to make digital humans accessible to everyone.
Digital humans have evolved through a 1.0 era of flat cartoon characters and a 2.0 era of virtual streamers into the 3.0 stage: high-precision models driven by artificial intelligence, now the industry mainstream.
One platform, full-chain tuning
The Baidu Intelligent Cloud XiLing platform is positioned as full-scene, one-stop and platform-based, covering everything from avatar creation and persona management to business orchestration, content creation and user interaction under multiple drive modes – meaning the whole chain can be tuned on one platform.
Li demonstrated the generation workflow: upload one photo to produce an AI-drivable 2D avatar quickly; a 3D digital human that once took two to three months can now be built in hours; and avatars can also be generated from a one-sentence voice description or simple face sculpting. Personalisation has improved too – Baidu’s newly released 3D ultra-realistic speaking avatar reaches 98.5% lip-sync accuracy.
Two classes of digital humans
By market use, digital humans split into performance type and service type. Performance type covers virtual anchors, virtual idols and digital-twin brand ambassadors; Baidu has built an AI sign-language anchor, CCTV.com’s virtual host Xiao C, a digital human for the Mars rover Zhurong, its own virtual brand ambassador Xijiajia and a digital version of celebrity Gong Jun.
Service type covers digital employees, virtual staff and virtual sales agents, bringing a new generation of customer service to finance, telecom operators, government and retail – easing long queues, slow responses and high labour costs. Baidu has partnered with SPDB, Everbright Bank and China Unicom on one-to-one service digital humans.
Three drivers behind the industry
Li names three core drivers: user demand and technology upgrades; policy support and capital inflows; and the iteration of computing platforms. Baidu’s breakthroughs rest on long-term positioning – it began a virtual human strategy back in 2019, and years of AI work are now the core capability of the XiLing platform, which the company is opening to the outside.
Want to talk through what you just read?
We work end to end, from product design to launch. Tell us about your scenario and we can start there.
Shenzhen XZS Technology Co., Ltd. · Tel 0755-82494862 · 13760424861