The former head of Google China said that China’s artificial intelligence models are 6 to 9 months behind the United States. | Wilnesh News

BEIJING – Kai-Fu Lee, the former head of Google China, said China’s artificial intelligence models may be at least half a year behind those developed in the United States, but China’s artificial intelligence applications may take off faster.

He’s referring to large language models that are trained on vast amounts of data to process and generate text, images, and videos.

Li said LL.M. programs at top Chinese firms lag about six to nine months behind their U.S. counterparts, while less advanced Chinese models may lag about 15 months behind the United States. He was speaking at the AVCJ China Private Equity Forum on Wednesday.

Li is the author of “AI Superpowers: China, Silicon Valley, and the New World Order,” a widely followed commentator on artificial intelligence, and the founder of startup 01.AI and venture capital firm Sinovation Ventures.

“I predict that by early next year, apps will become much faster in China than in the U.S.,” Li said, noting that the cost of training good artificial intelligence models has dropped significantly.

“China will inevitably (build) the best artificial intelligence applications in the world,” he said. “But it’s not clear whether it was built by a big company or a small company.”

Lee, whose startup is focused on search apps, said it could take five to eight years to take generative AI consumer apps to the next level — a “super app that can perform multiple tasks” “.

He said the industry may need entirely new devices compared to existing smartphones, adding that “the right device should be always on, always listening.”

Major Chinese companies such as Alibaba and Tencent Released their artificial intelligence models and commercial products. These companies and investors have also backed several AI startups.

Beijing-based Shenshu Technology, backed by Alibaba’s Ant Group, announced on Wednesday that its text-to-video model Vidu has launched a new feature for improving consistent depictions of key elements or characters in AI-generated clips , without distortion. This allows advertisers to create promotional videos for their products.

Vidu launched earlier this year, with its basic tools available to the public and more advanced features available via subscription. Co-founder and CEO Tang Jiayu told reporters on Wednesday that several companies are interested in buying Shengshu’s services and not just exploring the technology.