N
inety percent of the world’s digital content is inaccessible to 6.5 billion people — not because the ideas aren’t brilliant, but because they’re in the wrong language.
At TED, this keeps us up at night. Our mission has always been sharing transformative ideas, but what good are these ideas if most of the world can’t access them?
For more than 15 years, TED has relied on a network of 70,000 volunteer translators. Thousands of talks subtitled, in hundreds of languages, accumulating billions of views — an incredible effort. But anyone who’s struggled through subtitles knows the truth: Reading isn’t the same as experiencing.
A student in Brazil trying to keep up with subtitles. A commuter in Tokyo who prefers listening to reading. In many parts of the world, dubbed content isn’t just preferred — it’s expected. The demand for multilingual experiences was clear, and so were the challenges.
Advertisement
So, we moved beyond scaling subtitles to something more immersive — generative artificial intelligence (AI) that brings TED Talks to life in any language.
Partnering with Panjaya, a leader in AI video adaptation, we realized this was more than translation — it was a new way to transform content.
Imagine hearing Chimamanda Ngozi Adichie tell her story — not through subtitles, but in a voice that feels like she’s speaking directly to you. In Spanish, Arabic, or Japanese. Her rhythm, her emotion. Perfectly lip-synced.
The technical process works in three key steps:
But AI alone isn’t enough. Our expert-in-the-loop translators refine AI-generated dubs, ensuring every nuance and emotion stays true to the original. They also guard against AI hallucinations (misinterpretations that could alter meaning) so that what’s said in one language is just as powerful in another. This hybrid approach — AI plus human expertise — delivers an experience that feels authentic.
The result? A transformation in how global audiences experience TED Talks.
Advertisement
The Numbers Tell the Story
In 2024, we launched TED’s first AI-adapted multilingual talks in regions where dubbed content is widely preferred — Brazil, France, Germany, Spain, and Italy. The impact was immediate:
But numbers tell only part of the story. The real shift is in the response — speakers embracing the technology, audiences engaging in new ways, and AI-driven localization proving to be a game-changer for accessibility.
Ethical Guardrails: Innovation with Responsibility
Of course, we weren’t about to go all-in without guardrails. We built an ethical framework:
The Bigger Picture: The Future of Multilingual Content
The future of content is multilingual, and AI localization isn’t just a tool — it’s a strategic advantage. For TED, it’s the key to delivering a personalized, culturally relevant experience for a global audience.
As we scale this initiative to more than 20 languages and more than 5,000 adapted talks, our focus remains on authenticity and cultural accuracy, ensuring that knowledge flows freely — no longer confined by language.
This isn’t just about TED. It’s about how AI is transforming the way organizations, creators, and thought leaders connect with the world.
For us, the goal is simple: Ideas should be heard — by everyone, everywhere.
Helena Batt leads dubbing and localization at TED, making Talks feel native in 115 languages. With 15 years in product and language strategy, she helps millions access ideas that inspire — using AI, community, and a sharp focus on what it takes to truly connect across cultures.
Advertisement
Related Articles
The new features aim to enhance audio authenticity, improve workflow efficiency, and support both content creators and localization service providers in delivering high-quality projects on…
→ Continue ReadingThe new solution ensures impactful synthetic voice-over and dubbing using the latest speech synthesis technology.
→ Continue ReadingBeginning with ten languages, SYSTRAN will use TED content to develop neural machine translation models for technical content in a variety of fields. AI-based translation…
→ Continue Reading
