Artificial Intelligence 9 min read

Large Models Reshaping the Digital World – Speech by Baidu CEO Robin Li at the World Internet Conference Digital Civilization Dialogue

At the World Internet Conference, Baidu CEO Robin Li declared that large‑model AI, exemplified by Baidu’s upgraded Wenxin 3.5, is the new operating system driving digital transformation across industries—from automotive design to smart traffic—while emphasizing the need for native AI applications, safety safeguards, and robust regulatory frameworks to sustain economic growth.

Baidu Tech Salon
Baidu Tech Salon
Baidu Tech Salon
Large Models Reshaping the Digital World – Speech by Baidu CEO Robin Li at the World Internet Conference Digital Civilization Dialogue

On June 26, Baidu founder, chairman and CEO Robin Li delivered a speech titled “Large Models Reshaping the Digital World” at the “World Internet Conference Digital Civilization Ni Shan Dialogue”.

Li emphasized that large models are the current global focus of technological innovation and the main battlefield of the AI competition. He argued that the key to international competition is not the number of large models a country possesses, but the number of native AI applications built on those models and how much they improve production efficiency.

He announced that Baidu’s Wenxin large model has been upgraded to version 3.5, which doubles training speed, increases inference speed by 17 times, and improves overall model performance by more than 50% compared with version 3.0. The upgrade also brings significant safety enhancements, with noticeable gains in data quality, generation effect, and content security.

Li described the new AI technology stack as four layers: chips (GPU replacing CPU), frameworks (PaddlePaddle, PyTorch, TensorFlow), models (ChatGPT, Wenxin), and applications (AI‑native services). He positioned large models as the operating system of the AI era, with all downstream applications built upon them.

The speech highlighted how large models can deeply integrate with the real economy, citing examples in automotive manufacturing, energy, and transportation. In automotive design, large models can automatically generate design documents, shortening development cycles and reducing costs. In smart traffic, AI‑driven signal control can improve traffic flow, as demonstrated in Beijing’s Yizhuang district, achieving a 15‑30% increase in throughput and contributing to a 2.4‑4.8% GDP boost.

Li also warned of governance challenges, stressing the need for safety, responsible AI development, and the establishment of legal, regulatory, and ethical frameworks to ensure a healthy AI ecosystem.

He noted Baidu’s decade‑long investment in AI, its full‑stack layout across the four layers, and the company’s ability to perform end‑to‑end optimizations. Baidu’s smart traffic solutions have been adopted by 69 cities, and its AI cloud base in Shandong has fostered new AI‑related jobs and incubated 22 data‑annotation companies.

Concluding, Li asserted that large models will continue to permeate more sectors, driving digital transformation, creating substantial economic value, and reshaping the global digital economy. He called for a dual‑track approach of technological advancement and safety control to ensure sustainable progress.

He thanked the audience and ended his speech.

artificial intelligencelarge language modelsAI applicationsBaidudigital economy
Baidu Tech Salon
Written by

Baidu Tech Salon

Baidu Tech Salon, organized by Baidu's Technology Management Department, is a monthly offline event that shares cutting‑edge tech trends from Baidu and the industry, providing a free platform for mid‑to‑senior engineers to exchange ideas.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.