Artificial intelligence can reshape every industry in China, but only if it enhances productivity, said Gao Xuefeng, founder and CEO of Chinese tech firm Fabarta Co Ltd, at the 2025 Zhongguancun Forum in Beijing on Monday.

Gao said: "China's AI momentum, spurred by government support and market demand, is emerging as a key engine of economic growth. But AI must be transformed into productivity to have real vitality."

Fabarta, a proprietary multimodal knowledge engine and an industry-wide intelligent agent platform, aims to build a robust data infrastructure for the coming AI era.

Highlighting how past cycles of AI enthusiasm have faded when the technology failed to integrate deeply into real-world industries, he emphasized that the current wave of AI — driven by deep learning — is more resilient due to its increasing relevance with industrial operations.

To embrace such opportunities, Gao said the company is focused on a data-centric AI architecture rather than the traditional model-centric approach, which can better mitigate issues like model hallucinations and data security risks.

Model hallucinations are tendencies by large language models to sometimes concoct facts, invent unsolicited fictions or produce confident responses that belie an underlying falsehood.

Gao made the comments as Fabarta teamed up with Sinochem Information Technology Co Ltd and the College of Computer Science and Technology at Jilin University to establish a joint laboratory for AI and new materials.

The partnership, announced at the Zhongguancun Forum on Saturday, aims to embed AI into the full research and development cycle of new materials.

The lab will focus on four key areas — developing AI-powered vertical domain models to boost efficiency in materials research, building industry-specific data assets, enhancing intelligent tool platforms and integrating fundamental theoretical research with talent development.

Ge Shuang, Fabarta's co-founder and chief operating officer, said: "This joint lab is a testament to our commitment to using AI to drive industrial upgrades. Such a cooperation among three parties is expected to unlock AI's vast potential in materials science, foster an open and cooperative AI ecosystem and accelerate the transition to a smarter economy."

At this year's Zhongguancun Forum, several AI panels drew a packed house over the weekend, with attendees filling every seat and standing along the walls to listen. The overwhelming interest mirrored the prominence AI commanded, which could also be seen at the recently concluded Boao Forum for Asia Annual Conference 2025 and the China Development Forum.

As DeepSeek has become a popular topic at this year's forum, Gao said that after DeepSeek's breakthrough, many Chinese companies no longer hesitate to invest in AI.

"Every few months, new technological advances emerge. Companies can't afford to wait — they must integrate AI into their business and industry now," Gao said. "Every sector can be reimagined through AI in the years to come."

Gao added that in China's industrial AI landscape, large model deployment will likely begin with State-owned enterprises and major industry players, which have the necessary scale, infrastructure and data resources to invest in AI innovation.