Chinese developers take lead on using AI for good

Chinese developers are using artificial intelligence technologies to empower a wide range of sectors, unleashing its vast potential across art, traditional culture, ecological conservation, and education.

Several Chinese developer teams showcased their AI innovations at this year's Google I/O Connect China, which concluded on Aug 13 in Shanghai.

Zhao Yijie, a young artist graduating from the Central Academy of Fine Arts, aims to reintroduce ancient cultures into people's lives through art and technology. Using Google's Gemma 4 open model, he made it possible to interact with artifacts, pottery figurines, and paper cuts displayed in museums, rather than "observe" them.

Visitors can design their own paper cutouts by drawing on a tablet, then watch ancient pottery figurines mimic their dance moves while dressed in outfits with matching colors.

"We believe that the true preservation of traditional culture is not merely showcasing it in glass cases, but to invite audiences to step inside and become part of it, making traditional culture something they can experience, interact with, and connect with," Zhao said.

Han Peng, a postdoctoral researcher at the Institute of Ecology at Peking University, is working on biodiversity protection using the AI model Perch.

Home to more than 1,400 bird species, China is rich in avian biodiversity, resulting in a highly varied and complex array of bird songs that are extremely hard to monitor with traditional manual surveys. However, acoustic monitoring with AI offers a new solution to this challenge. Han noted that, compared with manual surveys, acoustic surveys may require less than 3 percent of the time and incur a much lower cost to achieve the same survey objectives.

"AI's ability to understand the meaning behind bird songs goes beyond simply identifying a bird species," said Han. He hopes to use AI to build a database of Chinese bird songs. "Its greater value lies in helping us hear the signals ecosystems emit and ultimately turning these sounds into scientific evidence that can help us understand changes in nature and protect biodiversity."

Luo Wencong, a visually impaired software engineer at Google, focused on developing and improving screen reader features for his community.

NVDA is an open-source screen-reading software launched 20 years ago by two visually impaired Australians. Luo, along with other volunteers, has established the software's local community in China and contributed new features to the project.

One challenge that has long troubled its users is that when the reader encounters an image, it cannot interpret its content. To tackle this problem, Luo's team developed VisAware. This visual assistance plug-in can help with reading book illustrations and complex formulas, browsing social media content, and performing complex interactions.

"Since the plug-in's release, we've received a lot of genuine feedback and positive reviews from users with visual impairments, which makes me particularly happy," Luo said. "As a developer, I've always been motivated to contribute when I'm able to use technology to tangibly help users solve specific challenges in their daily lives and work."

Source: China Daily