By AI Trends Staff

The world of speech recognition is rapidly evolving, fueled by advances in artificial intelligence and machine learning. This evolution is propelling the market forward, capturing the interest of venture capitalists, and challenging established players.

The acceptance and utilization of speech recognition technology have been gaining momentum. According to a report by Meticulous Research, the global speech recognition market is expected to reach $26.8 billion by 2025. Speed and accuracy are among the key advantages driving this growth.

The Genesis of a Speech Recognition Leader

Founded in 2017 by Dylan Fox, AssemblyAI has been a trailblazer in making speech recognition technology accessible and accurate. Dylan’s journey from a business graduate to a top-tier tech entrepreneur is both fascinating and inspiring.

Fox, armed with a degree in business administration, business economics, and public policy from George Washington University, initially started as a software engineer focusing on machine learning at Cisco. It was here, working on deep neural networks and machine learning, that he conceived the idea for AssemblyAI.

The Visionaries behind AssemblyAI

During an interview with AI Trends, Fox recounted his transition from a business undergrad to a high-tech entrepreneur. “I taught myself how to program, which led me to a path of machine learning. I was looking for a harder software challenge, which led to natural language processing, which took me to Cisco,” said Fox.

At Cisco, Fox had a front-row seat to the search for speech recognition software. “We examined options like Nuance, which was recognized as a market leader. It was surprising how inadequate these solutions were, both in terms of accuracy and developer usability,” Fox revealed.

Revolutionary Approach to Speech Recognition

Fox drew inspiration from Twilio, a company that set new standards for API development. His mission was clear: use AI and machine learning to achieve super accurate results and simplify the integration of the API for developers.

AssemblyAI attracted marquee clients such as CallRail, NBC, and the Wall Street Journal, who use their API for accurate transcription, closed captioning, and other services. “We’ve been working towards building speech recognition quality as close to human as possible,” Fox said. He anticipates reaching this milestone in 2022.

Flexible, Scalable, and Developer-Friendly

AssemblyAI’s model targets companies incorporating speech recognition into their products. Clients benefit from a pay-as-you-go model, making costs predictable and scalable. For every second of audio transcribed, clients are charged a fraction of a penny, billed monthly. This translates to about nine dollars for 10 hours and $900,000 for a million hours of audio transcription.

Advanced AI Features

AssemblyAI’s capabilities go beyond mere transcription. Their technology identifies sensitive topics like hate speech and profanity, offering significant savings on human content moderation. Additionally, they build AI features on top of transcriptions, providing searchable, indexed summaries of audio and video content.

“We are an experienced team of deep learning researchers,” Fox explained, “with expertise from tech giants like BMW, Apple, and Facebook. We build large, accurate deep learning models using cutting-edge neural network technology.” This methodology is akin to OpenAI’s approach with its GPT-3 language model.

Addressing the Future

AssemblyAI is poised for significant growth, with plans to double its current 25-employee workforce within four months. “There is an explosion of audio and video data online, and customers want to leverage it,” noted Fox. The demand for their services is robust and escalating.

For more information, visit AssemblyAI.