Microsoft introduces low cost speech recognition model

Microsoft announced a new artificial intelligence system called MAI Transcribe 2, designed to convert spoken audio into written text. According to an announcement from the tech giant, the tool handles 60 languages and includes features like identifying different speakers and placing exact time stamps on individual words.

For everyday users, this means automated transcription services could soon become drastically cheaper and more accessible across consumer apps. Microsoft is charging ten cents per hour of audio processed, significantly undercutting older models. Lower costs make it much easier for developers to add fast, reliable closed captioning and meeting summaries to software without charging high subscription fees.

Microsoft claims the software outpaces rival offerings from competitors on standard benchmark tests for accuracy and speed. By making audio conversion cheaper and cleaner, tools that rely on voice input, such as virtual assistants or accessibility features for the hard of hearing, could see noticeable performance gains.

It remains unclear how well the software performs in noisy real world environments outside of lab benchmarks, or how privacy will be managed when processing sensitive personal conversations. Additionally, while lower pricing benefits software builders, it raises questions about whether smaller competing AI startups can survive such aggressive price cuts from tech giants.