I still remember the first time I tried to use voice recognition on my computer back in the late 90s. It was a disaster. I spent hours training the software, speaking like a robot, only for it to produce complete gibberish. I think it transcribed “Hello, world” as “Hell, word”. I gave up, convinced it was a gimmick that would never work.
Fast forward to today, and I’m having full-blown conversations with the AI in my car, my phone, and even my glasses. My kids are growing up in a world where talking to computers is as normal as typing on them. As someone who has been in the tech world for over two decades, with a couple of successful exits and over 200 angel investments in companies like Anthropic and OpenAI, I’ve had a front-row seat to the AI revolution. And let me tell you, AI voice is one of the most exciting and disruptive parts of it.
But what is AI voice, really? And why is it suddenly everywhere? Let’s break it down.
The Magic Behind the Curtain: How AI Voice Actually Works
At its core, AI voice technology is about two things: teaching computers to understand what we say, and teaching them to talk back to us. It sounds simple, but it’s taken decades of research and billions of dollars in investment to get to where we are today.
Speech Recognition: From “Hell, Word” to Flawless Transcription
Speech recognition, or Automatic Speech Recognition (ASR) as it’s known in the industry, is the technology that converts spoken words into text. The early systems, like the one I used in the 90s, were rule-based. They had a very limited vocabulary and struggled with different accents, background noise, and even just a slight deviation from the “correct” way of speaking.
Today’s ASR systems are powered by deep learning, a type of AI that learns from massive amounts of data. Think of it like this: you can’t teach a child to understand language by giving them a grammar book. They learn by listening to people talk, day in and day out. That’s essentially what we do with AI. We feed it millions of hours of audio data – everything from audiobooks and podcasts to phone calls and YouTube videos. This allows the AI to learn the nuances of human speech, including different accents, dialects, and speaking styles.
This is why the AI voice assistant on your phone can understand you whether you’re in a quiet room or a noisy street. It’s also why we now have services that can transcribe meetings and interviews with incredible accuracy, saving us hours of manual work.
Text-to-Speech: Giving AI a Voice
Once the AI understands what you’ve said, it needs to be able to respond. That’s where Text-to-Speech (TTS) comes in. TTS technology converts written text into spoken words. Like ASR, early TTS systems were very robotic. They sounded like a computer reading text, with unnatural intonation and rhythm.
Modern TTS systems, however, are a world apart. They use generative AI models to create voices that are almost indistinguishable from a real human. These models can even be trained to replicate a specific person’s voice, a technology known as voice cloning.
Voice Cloning: The Power and the Peril
Voice cloning is one of the most fascinating and controversial areas of AI voice. On the one hand, it has the potential to do incredible good. For example, it can give a voice back to people who have lost their ability to speak due to illness or injury. It can also be used to create more personalized and engaging experiences in everything from gaming to education.
I’ve personally invested in a few companies that are pushing the boundaries of what’s possible with voice cloning. One of them is developing a technology that can create a realistic voice clone from just a few seconds of audio. The potential applications are mind-boggling.
But on the other hand, voice cloning also opens up a Pandora’s box of ethical concerns. What happens when you can no longer trust what you hear? We’ve already seen the rise of deepfake videos, and deepfake audio is not far behind. Imagine getting a call from your boss asking you to transfer money, but it’s not really your boss. It’s a scammer using a voice clone. That’s a scary thought, and it’s something we as a society need to be prepared for.
Where I’m Placing My Bets: The Most Exciting Applications of AI Voice
As an investor, I’m always looking for the next big thing. And right now, I’m incredibly bullish on AI voice. I believe we’re on the cusp of a major shift in how we interact with technology, and voice is at the center of it. Here are a few of the areas where I see the most potential:
AI Voice Assistants That Don’t Suck: Let’s be honest, most of the voice assistants we use today are still pretty dumb. They can set a timer or tell you the weather, but they struggle with complex conversations. That’s about to change. The next generation of voice assistants will be truly conversational. They’ll be able to understand context, remember past conversations, and even anticipate your needs. I’m not talking about just another smart speaker. I’m talking about a true digital assistant that can help you manage your life, both personal and professional.
The Future of Customer Service: Nobody likes calling customer service. You have to navigate a maze of automated menus, wait on hold for what feels like an eternity, and then repeat your problem to multiple different people. AI voice is set to completely transform this experience. Companies are already using AI-powered voice agents to answer common questions, route calls to the right department, and even handle complex issues. This not only improves the customer experience, but it also frees up human agents to focus on the most challenging problems.
Hyper-Personalization at Scale: AI voice allows for a new level of personalization that was never before possible. Imagine a world where your favorite news podcast is read to you in the voice of your choice. Or where the characters in a video game can have a unique voice that’s generated on the fly. This is the kind of hyper-personalized experience that AI voice can deliver, and it’s going to change the way we consume content.
The Dark Side: Navigating the Risks of AI Voice
As with any powerful technology, AI voice comes with its own set of risks. I’ve already touched on the dangers of voice cloning, but there are other concerns as well.
One of the biggest is privacy. For AI voice to work, it needs to be always listening. That means our smart speakers, our phones, and even our cars are constantly collecting data about us. The question is, what happens to that data? Who has access to it? And how is it being used? These are important questions that we need to be asking.
Another concern is the potential for job displacement. As AI voice agents become more capable, they will inevitably replace some of the jobs that are currently done by humans, particularly in areas like customer service. This is a real concern, and it’s something we need to be proactive about addressing. We need to invest in education and training to help people transition to the new jobs that will be created in the AI economy.
My “Top 1%” Take: What’s Next for AI Voice
So, what’s my advice for founders and investors who are looking to get into the AI voice space? Here are a few of my predictions for the next 3-5 years:
The Rise of Proactive Voice Assistants: The next generation of voice assistants will be proactive, not reactive. They’ll anticipate your needs and offer suggestions before you even ask. For example, your car’s voice assistant might notice that you’re low on gas and suggest a nearby gas station. Or your phone’s assistant might remind you of an upcoming meeting and ask if you want to review the agenda.
The Blurring of the Lines Between Human and AI Voices: As TTS technology continues to improve, it will become increasingly difficult to tell the difference between a human voice and an AI voice. This will open up new possibilities for creative expression, but it will also make it more important than ever to have safeguards in place to prevent misuse.
The Emergence of a “Voice-First” Internet: Just as we’ve seen a shift from desktop to mobile, I believe we’re going to see a shift to a “voice-first” internet. More and more people will be using their voice to search the web, shop online, and interact with their favorite apps and services. This will require a fundamental rethinking of how we design and build digital experiences.
The Future is Spoken
AI voice is no longer a gimmick. It’s a powerful technology that is already changing the way we live and work. And we’re still in the very early days. The opportunities for innovation are massive, and I for one am excited to see what the future holds.
But as we embrace this new technology, we also need to be mindful of the risks. We need to have a public conversation about the ethical implications of AI voice, and we need to put safeguards in place to ensure that it’s used for good.
The future is spoken. The only question is, what will it say?
Frequently Asked Questions
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.