When you think of voices in games, what comes to mind? Is it the rigid voice actor who only ever says "Hello, I'm an NPC," or is it the situation where, due to insufficient development funds, all the dialogue in the game relies on subtitles?
On stage at this year's Taipei Game Show, Chao Hsu-Kuang, founder of VoAI, stated frankly: "Today's AI voices are no longer the mechanical voices of the past." This is not just a technological advancement, but also a liberation of "productivity" and "imagination."
When "voice-over" is no longer a "luxury"
At the beginning of his speech, Chao Xuguang pointed out an interesting phenomenon: in the age of information overload, people's eyes are filled with all kinds of short videos and messages, and "hearing" has become a new battleground for filling fragmented time. From driving, doing housework to commuting, the ears are idle for a lot of time.
However, for game developers or content creators, producing high-quality audio content is often prohibitively expensive. Not only do they need to hire professional voice actors, but they also need to rent recording studios, and post-production editing is even more costly. As a result, many games (especially indie games) are forced to abandon voice acting and opt for text-only presentation.
"If there were a robot today that could speak like a real person, using your voice, your boss's voice, or even the voice of an internet celebrity, that would be really interesting." Chao Xuguang emphasized that VoAI's core technology is to make AI voice-overs more than just the monotonous "Miss Google" with its varied voices; it can now convey a wide range of emotions (happiness, embarrassment, anger) and even mimic specific character personalities (such as young women, young men, or storytellers).
Chao Hsu-kuang further pointed out that although there is already Chinese audio data available globally, most of it is still training content for the Chinese market. The Taiwanese market generally needs audio data that is more localized, so training content that meets the needs of the local market is expected to be more valuable.
AI NOWCAST: Your LINE is your AI news station
Beyond simply providing voiceovers, VoAI also showcased another of its practical applications – AI NOWCAST, an automated news broadcasting platform built on LINE.
Its operating logic is very simple: it captures news text, rewrites it into a script suitable for spoken broadcast through AI, generates audio content, and finally pushes it to users to listen to. Behind this is actually the demonstration of AI's information processing capabilities, including "structuring, summarizing, and re-speech".
For the gaming industry, the potential of this technology extends far beyond simply "broadcasting news." If this logic is applied to in-game "daily quest briefings" or "guild battle updates," players no longer need to stare at densely packed announcements. Instead, they can directly "hear" an AI character telling them in a vivid tone, "Guild Leader! Our base was attacked last night, and we suffered heavy losses!" This would significantly enhance the sense of immersion.
AI avatars: They not only need to speak, but also have a "face".
In addition, Chao Xuguang also showcased VoAI's latest technology at the event—AI video cloning. With just a photo and a voice recording, AI can make the person in the photo "move" and speak, even matching their lip movements.
For game developers, this technology offers a significant advantage over traditional methods. Creating a moving, talking NPC previously required 3D modeling, motion capture (Mocap), and facial expression rigging, each step of which incurred costs. With AI technology, a single 2D artwork, combined with AI-generated voice, can quickly produce a large volume of narrative content.
Chao Xuguang further proposed the concept of "mass-produced video content," which uses AI Agents to train AI to target different audiences and automatically generate thousands of different short videos. For game marketing, this means that a massive amount of advertising materials can be tested at extremely low cost, and the one with the highest user conversion rate can be found.
This may also connect to more market opportunities in the future.
Chao Hsu-kuang stated that VoAI currently offers services in two main categories: SaaS (Software as a Service) and API access. It is primarily built on the Google Cloud platform, but also collaborates with AWS, Azure, and local cloud service platforms in Taiwan to mitigate traffic and system operation risks.
And KK Company, built with blockchain and AI technologiesTheKeeper, an AI-powered digital avatar management platformIn contrast, Chao Xuguang explained that VoAI actually possesses similar technology and is also considering the possibility of developing in conjunction with the entertainment market, but at this stage, it mainly collaborates with creative groups such as YouTubers and KOLs.
Analysis: AI is an amplifier of creativity, not a substitute.
Chao Xuguang emphasized that AI voice technology is filling the gaps in game development where "we wanted to do it but couldn't".
Often, developers don't lack the desire for full voice-over, but rather the budget; they don't lack the desire for dynamic presentations, but rather the manpower. The emergence of generative AI, however, transforms these "luxuries" into "standard" technologies.
Of course, AI voices may not yet match the emotional nuances of top professional voice actors. However, in terms of quantity and speed, AI has already demonstrated absolute capability. The future game development process may become: the core storyline is performed by real voice actors, while a massive amount of side quests, random dialogues, and even dynamically generated NPC interactions will be entirely handled by AI.
When the cost of generating sound and images approaches zero, all that remains is for game developers to test their creativity and storytelling abilities.
