How Microsoft’s First Chinese AI Fellow Is Driving Speech and Language Breakthroughs

Microsoft announced that Huang Xuedong, the company’s first Chinese Global Technical Fellow and former chief speech scientist, has been appointed Global AI Chief Technology Officer, responsible for Azure AI services including speech, machine translation, computer vision, natural language, and cognitive services.

Huang entered Hunan University’s teacher training program at age 15, earned a bachelor’s degree there, completed a master’s at Tsinghua University, and received a PhD from the University of Edinburgh. He joined Microsoft Research in 1993, built the speech recognition team, became chief speech scientist, and is one of Microsoft’s 20 global technical fellows.

His team achieved a historic milestone in 2016 by reaching a 5.9% word error rate (WER) on the Switchboard benchmark, matching professional stenographers. In 2017 the WER improved to 5.1%, surpassing human performance. In 2018 Microsoft’s machine translation system achieved human‑comparable quality on Chinese‑English news translation, and in 2019 the reading‑comprehension system outperformed humans on the Stanford dialogue QA leaderboard.

In an interview, Huang outlined three key focus areas: (1) advancing AI capabilities by improving speech recognition accuracy, enhancing speech synthesis naturalness, and ensuring translation precision; (2) delivering these technologies as modern cloud and edge services through Azure, including SDKs for devices; and (3) fostering strong customer relationships to provide first‑class experiences.

He emphasized that "speech and language are the crown jewels of AI," arguing that mastering these core technologies is essential for true artificial intelligence progress.

Huang also discussed how speech, translation, and dialogue enable AI to evolve from perception to cognition, bringing machines closer to human‑like intelligence.