PII Redaction for Speech-to-Text Transcriptions (March 2020 Update)

AssemblyAI's research team has introduced a new neural network update that significantly improves accuracy and speed. The asynchronous English model consistently outperforms competitors in terms of accuracy, with an average Word Error Rate (WER) lower than Google Cloud's video model, AWS Transcribe, and Microsoft Azure. Additionally, the platform now offers PII redaction to automatically detect and remove sensitive information like credit card and social security numbers from transcriptions. Furthermore, the Speed Boost feature enables faster transcription of short audio files (1 minute or less) by completing them 25-50% faster than usual. AssemblyAI has also maintained a 100% uptime across all models for another month.