With the advancement of voice artificial intelligence (AI), language barriers will be completely broken down in the future. Real-time dubbing and translation will make communication possible anywhere in the world.
Co-founder and Chief Executive Officer (CEO) Mati Staniszewski of British artificial intelligence (AI) audio company ElevenLabs said this as the company officially announced its entry into the Korean market at a press briefing held on the 21st at the JW Marriott Hotel in Seocho-gu, Seoul.
CEO Staniszewski said, The goal of ElevenLabs is to give technology a voice, adding, Speech is the most primal form of communication, and the voice carries fine emotions that text cannot fully express, so in the future, when we call customer service, drive cars, use smartphones, and further, communicate with robots, voice will be at the center of it all.
Founded in London on the 21st, ElevenLabs is a company that uses AI-based Text-to-Speech (TTS) technology to render voices in various languages. It has technologies such as TTS (Text-to-Speech), voice cloning, AI dubbing, and sound effects that convert text into a human voice in real time.
At the briefing, the company demonstrated a greeting in Korean generated by AI cloning CEO Staniszewski's voice. I do not speak Korean, he said, but using ElevenLabs' voice cloning technology, I can render my voice speaking fluent Korean. It is my voice, but it feels hyperreal because it is so identical.
ElevenLabs' technology has about 50 million monthly users, and in three years since its founding, its valuation has jumped to $6.6 billion (about 9.7 trillion won). CEO Staniszewski said, Seventy-five percent of Fortune 500 companies are customers, adding, In Korea, leading corporations such as Naver, LG Uplus, and Krafton Inc. are using ElevenLabs' technology. Nvidia, Deutsche Telekom, U.S. actor Matthew McConaughey, and LG have invested in ElevenLabs.
CEO Staniszewski, who is from Poland, and co-founder and Chief Technology Officer (CTO) Piotr Dabkowski said they started ElevenLabs after feeling frustrated that Polish dubbing of imported films always used the same voice and becoming interested in AI-based dubbing technology to solve that. The experience of one voice actor dubbing the voices of all characters in imported films was truly terrible, CEO Staniszewski said, adding, We started with dubbing, but we saw great potential for broader advances in the audio technology space and expanded the business.
ElevenLabs' flagship AI agent platform converts typed text into more than 7,000 voices and 32 languages in 0.5 seconds.
He said, To enter the Korean market this time, we invested significant resources by hiring experts and forming a dedicated team to develop models that generate Korean text and speech, adding, We worked to accurately capture and render Korean-specific pronunciation, intonation, and emotion. ElevenLabs plans to apply this technology to Korea's content and game industries and use it to improve experiences at customer centers where AI transitions are happening rapidly.
With this, ElevenLabs became the fifth AI unicorn (a private company valued at 1 trillion won or more) this year to establish a corporation or office in the Korean market. Previously, OpenAI, known for ChatGPT, Claude developer Anthropic, Cohere, and Anduril entered the Korean market.
Like other AI companies, ElevenLabs said it decided to establish its sixth office in Seoul, noting the rapid growth of the Korean AI market. Hong Sang-won, head of ElevenLabs Korea, said, Korea is the fastest market to embrace innovation, adding, Sixty-five point one percent of large corporations have already introduced AI, and 63.5% of workers use Generative AI in their daily work, which is more than double the global average.
He also said, Korea's global content power proven by K-pop and K-drama, and the world's most demanding service standards make the Korean market optimal, adding, Success in Korea is a barometer of global success, so ElevenLabs chose Korea as its key bridgehead for entry into Asia.
In the domestic market, the company said it will focus on globalizing K-content and help Korea leap forward as an "Asian voice AI hub." Hong said, K-content has captivated the world, but language barriers still exist, adding, ElevenLabs' "Eleven v3" model supports more than 70 languages and almost perfectly reproduces the original emotions and nuances, helping remove this barrier. It conveys not just translation but also laughter, sighs, exclamations, and even breathing sounds as they are.
AI voice products using the voices of Korean celebrities are also in the works. Early this month, ElevenLabs unveiled Iconic Voices, which commercially sells AI-cloned voices of celebrities. Users can access the voices of about 30 famous people, including historical figures like Thomas Edison and Alan Turing, actor Michael Caine, and baseball players. A key feature is formalizing licensing agreements so corporations can legally use celebrity voices. CEO Staniszewski said, Korea has a wealth of talent in K-drama and K-pop, adding, We also look forward to collaborations with Korean celebrities.
He emphasized that ElevenLabs' technology can also help increase the efficiency of customer service. Ultra-low-latency voice AI agents with response times under 500 milliseconds can respond to customer inquiries around the clock in multiple languages, reducing agents' workloads.
CEO Staniszewski predicted that as voice AI technology advances, a time will come when every device—from smartphones and home appliances to cars—understands human speech along with its context and emotions and interacts accordingly. All devices will have "voice intelligence" that understands even human speech patterns, he said, adding, Real-time translation and dubbing technologies will remove language barriers so that even when traveling abroad, people can understand surrounding information in the local voice and intonation.
ElevenLabs is also preparing for an initial public offering (IPO). He said, We set a goal to list within five years, but if the current growth continues, it could be possible within three years.