Alibaba's Joseph Tsai Discusses AI Development in Wuzhen: Only Open Source and Openness Can Bring Inclusive Access

On the morning of November 3, the 2026 World Internet Conference Wuzhen Summit opened. Joseph Tsai, co-founder and chairman of the board of Alibaba Group, spoke at the opening ceremony, saying that AI can only become widespread if it is open source, and only widespread adoption can create a sustainable, inclusive intelligent era. China's open-source models remain ahead on key indicators such as download volume and the number of derivative models, making the country the current global leader in open-source intelligence.

(Photo shows Joseph Tsai, co-founder and chairman of the board of Alibaba Group, speaking at the opening ceremony of the Wuzhen Summit.)

Tsai said that over the past year, global AI development has accelerated, pushing this productivity transformation to greater depth. AI use cases have continued to diversify and scale, extending into fields such as industrial manufacturing, healthcare, public services, enterprise office work and scientific research, with AI demand continuing to climb and computing power in short supply. Chinese technology companies have continued to step up model open-sourcing and ecosystem openness, remaining ahead on key indicators such as open-source model downloads and the number of derivative models, and becoming the current global leader in open-source intelligence. Chinese open-source models offer strong overall performance and high inference efficiency, are widely praised around the world and are highly competitive.

Tsai believes that Chinese AI companies independently chose open source because of the core spirit of Chinese innovation: rising above self-interest to advance the common good, promoting the enhancement of wisdom through sharing, and being willing to create greater value by helping each other succeed. He said: "Only open source can make AI widespread, and only widespread adoption can create a sustainable, inclusive intelligent era."

Open source and openness: lowering R&D costs for SMEs and benefiting developing countries

Three years ago, Alibaba proposed an "open source and open" AI strategy. To date, Alibaba's Qwen has open-sourced more than 470 models, with cumulative global downloads exceeding 4 billion and more than 330,000 derivative models, making it the world's No. 1 open-source model family. Tsai said open source promotes the prosperity of the global developer ecosystem, effectively lowers R&D costs for SMEs, and is especially valuable for low-resource languages and developing countries. Based on open-source models, AI is no longer the exclusive preserve of a few technological powers but a cause in which all parties can participate deeply. Developers from the vast emerging economies can quickly build their own national AI applications and innovation ecosystems without starting from scratch, effectively bridging the intelligence gap between countries and regions. He gave an example: in Uganda, Africa, a nonprofit organization developed a local-language model based on the Qwen open-source model, which now covers 68 African languages, helping more and more Africans overcome the threshold of foundation models and break through language barriers to step through the door of AI together.

According to the introduction, ModelScope, created and operated by Alibaba, is China's largest AI open-source community and has attracted 33 million users from 239 countries and regions around the world. In this open community, global developers can access 280,000 open-source models, 40,000 datasets and 12,000 MCPs (Model Context Protocols). Tsai said: "Rooted in China and serving the world is the positioning of the ModelScope community and also Alibaba's mission for AI ecosystem building."

At this World Internet Conference, Alibaba showcased the latest achievements of its T-Head chips, Alibaba Cloud's "cloud + AI" services, the Qwen large models and a series of AI applications. Alibaba's agent platform Qoder won the "Leading Technology Award" for its technical breakthroughs in the agent Harness, making it the only general-purpose agent product in China to receive the award.

In the medical field, the "plain CT + AI" technology launched by Alibaba DAMO Academy has continued to make breakthroughs, becoming the first in the world to enable simultaneous screening for five types of cancer, two types of emergencies and five types of chronic diseases in a single plain CT scan. DAMO RADAR, a general medical imaging AI model recently developed jointly by DAMO Academy and institutions including the First Affiliated Hospital of Zhejiang University School of Medicine, can identify more than 146 types of diseases at once, with accuracy reaching the level of imaging specialists for the first time; the results were published in the top international academic journal Science. Alibaba open-sourced this model immediately.

In the consumer field, Alibaba's Qwen App took the lead this year in enabling "AI errands," providing users with services such as food ordering, shopping and flight and hotel bookings, and recently launched a dedicated personal AI assistant, "Qwen Go!," which can handle more types and more areas of personal affairs.

In the office field, Alibaba recently launched a one-stop AI office platform, "Qwen Office," deeply integrated into the DingTalk ecosystem and connected with enterprise databases and mainstream instant messaging tools, integrating multimodal content generation, browser automation, web page creation, scheduled task scheduling and a full set of Office document processing capabilities to help enterprises improve organizational efficiency.