Artificial intelligence is emerging as an unexpected ally in the effort to preserve Cantonese, a language under pressure from the growing dominance of Mandarin and the shrinking number of young speakers willing to learn it. In Hong Kong, one deep-tech company is trying to use large language tools to protect Cantonese and the cultural identity tied to it.
What Happened
The challenge facing Cantonese is structural as well as cultural. Mandarin has become the dominant language across much of China, while Cantonese has struggled with fewer learning resources and the absence of a fully standardized written form. Those obstacles have made it harder for the language to compete in schools, workplaces and digital spaces.
That pressure has left Cantonese with an uncertain future, especially among younger generations. As fewer young people choose to study or use it regularly, the language risks losing everyday relevance even in places where it remains deeply rooted in local identity.
Votee AI, a Hong Kong-based deep-tech company, is trying to reverse that trend by applying AI to language preservation. The idea is to build tools that can help keep Cantonese usable, searchable and teachable in modern settings where technology increasingly shapes how people communicate.
Background
Cantonese has long been more than a spoken language. It is closely tied to Hong Kong’s history, media, cinema, music and civic identity, and it remains widely spoken in southern China and across parts of the global Chinese diaspora. But the rise of Mandarin as the standard language in education, government and national media has steadily changed the balance of power.
Language preservation has become a broader global issue as communities confront the loss of regional languages and dialects. Once a language loses daily use among younger speakers, it can be difficult to rebuild. Schools can help, but digital tools often determine whether a language feels modern enough to survive in daily life.
That is where AI may matter most. Language models, speech tools and translation systems can lower the barrier to learning and using less-dominant languages. They can help create learning materials, improve transcription and make it easier to interact with written and spoken forms that may not have extensive digital archives.
Why It Matters
The effort to preserve Cantonese is about more than vocabulary or pronunciation. It is about whether a major regional language can remain visible in an era when digital infrastructure increasingly rewards standardized, high-volume languages. If AI can help sustain Cantonese, it could offer a model for protecting other endangered or marginalized languages around the world.
For readers in Panama and Latin America, the story also echoes a familiar concern: how to preserve linguistic and cultural diversity in a globalized digital economy. Indigenous and regional languages across the Americas face similar pressures from dominant national languages and limited educational resources. If AI tools can be adapted for preservation, they may become part of a wider effort to keep local identities alive.
The broader significance extends beyond language itself. In an age of rapid technological change, the tools that shape communication can either flatten cultural differences or help defend them. For Cantonese, the outcome may depend on whether innovation reaches the communities that need it most.