Artificial intelligence unlocks silver screen for people with disabilities

The screening hall fell quiet as the film began. Soon, a steady narration described the gestures, expressions and moments that could not be seen by people with visual impairments. For them, cinema unfolded not through images, but through language.

That calm, precise and unmistakably familiar voice belonged to Chinese actor Hu Ge, recreated through artificial intelligence.

"I may not be able to go offline to tell stories in person to my blind friends," Hu said in a recorded message. "But my voice is online. Whenever they want to watch a movie, they can come and listen to me tell it."

The project sits at the center of Youku's accessible theater, an initiative that uses AI to transform mainstream film and television into formats accessible to visually impaired, hearing-impaired and elderly audiences. Behind it is Chen Yanling, head of public welfare at Alibaba Group's Hujing Digital Media and Entertainment Group.

"People used to think voice packs were only for navigation or commercial use," Chen said. "I never imagined that one day blind people could watch movies through a voice pack — and that it would be Hu Ge telling them the story."

Chen, now in her 23rd year at Alibaba, did not begin with technology. She began as a volunteer.

In 2022, Chen was providing live audio descriptions at offline accessible film screenings in Beijing. Because the events were held in person, many participants traveled long distances. Some were in their 60s, coming from remote districts such as Changping and Pinggu.

"When I walked them to the subway, they told me that as long as there was accessible cinema, distance didn't matter," Chen recalled. "Even if they had to leave home at four or five in the morning, they would come."

The experience stayed with her.

At the time, Chen had just transferred back to Youku, Alibaba's video platform. She began asking whether technology could remove those physical barriers — and whether an online platform could make accessible cinema part of everyday life.

"Technology focuses on efficiency and standards. Entertainment focuses on creativity and copyright," she said. "But public welfare focuses on warmth and sustainability. The challenge was how to combine all three."

She began coordinating informally across teams — engineers, copyright specialists and operations staff — asking whether they could create an online space for accessible films. She also volunteered to narrate films herself, drawing on years of experience as an audio describer.

Within a week, Youku launched its accessible theater with three films. The response exceeded expectations. "The viewership was even higher than some very popular movies at the time," Chen said.

But success quickly revealed limits. Manual narration was slow and labor-intensive, and copyright restrictions made free access unsustainable. This was where AI became essential.

According to materials released by Youku, a two-hour film traditionally required at least three days to convert into an accessible version. With AI-assisted processing, that time can now be reduced to about two hours. The technology helps generate descriptive scripts, align narration precisely with visuals and speed up production without altering the original meaning of the content.

"This efficiency is what made scale possible," Chen said.

From three films, the catalog expanded to 100, then 2,000, and today, more than 9,200 accessible titles. Rights holders agreed to participate on a trial basis. "They told us: if the original meaning changes, we stop," Chen said. "So we were very careful."

AI also enabled expansion beyond visual impairment.

In March, Youku opened identity authentication for hearing-impaired users, allowing free access to accessible content. Instead of relying on sign language, the platform uses AI-assisted subtitles that identify speakers and annotate background sounds — such as piano music, wind or tension-building effects — elements that are crucial to storytelling.

"In many films, background sound is essential to emotion," Chen said. "Even fear comes from sound."

An official from the China Disabled Persons' Federation said technological progress was expanding the possibilities of inclusion. "With the rapid development of AI, equality, integration and sharing are becoming more achievable for people with disabilities," the official said.

Zhang Jiantong, a university student who is hearing-impaired and among the first certified users, said the platform has become part of her daily routine. "It feels very good," she said. "I watch for more than an hour in the evenings. It gives me more topics to talk about with classmates and family."

AI has also been applied to Youku's "Silver Hair Theater", which adapts content for elderly viewers. Large-font subtitles, background noise reduction and specially developed "elder-friendly audio" help address common problems such as unclear dialogue and improper volume levels.

Chen said the idea came from her own family. "My father is more than 70 and can't hear clearly," she said. "The TV volume was so loud I couldn't even walk into the room."

For Chen, AI is not only about speed or efficiency. "You must embrace AI and accept the arrival of this era," she said. "But technology is not just about efficiency and speed, it is also about 'temperature'."

At the end of 2025, Chen experienced what she called a "life highlight" when she carried the torch at the Milano Cortina 2026 Winter Olympics. "This moment was not only a personal honor," she later wrote in her WeChat moment. "It was recognition for insisting on doing the right and meaningful things."