AI-assisted speech transcription for parliamentary proceedings

Transcription and translation

Canada - House of Commons

Use case ID: 067 
Author: House of Commons of Canada 
Date: 17 July 2024 

Objective:

Enhance the accessibility of parliamentary debates and committee meetings by implementing artificial intelligence (AI)-assisted speech recognition technology, thereby producing a live transcript directly from the audio feed from the room, with delays as close as possible to the Canadian Radio-television and Telecommunications Commission (CRTC) standards.

Actors:

  • Parliamentary digital media staff preparing the event
  • AI systems for automated speech recognition
  • Parliamentary information management system, including digital transcription and authoring modules
  • Live transcription application
  • Event participants: MP(s), committee witness(es), visitor(s) and interpreter(s)

Prerequisites:

  • Digital audio live feed of parliamentary proceedings
  • Automated AI speech recognition technology
  • Metadata to identify speakers and languages within an audio segment; alternatively, speaker diarization and language detection technologies can be used
  • Integration with parliamentary digital transcription authoring systems and workflow
  • Integration with the Live Transcription App

Scenario:

  • A participant in a parliamentary plenary session or committee meeting speaks.  
  • The audio captured during the speech is routed to the AI system.
  • The AI system detects the language being spoken and generates the transcript.
  • A participant is using a managed device with the Live Transcription App that is receiving the generated transcript.

Alternate flows:

  • When the automated AI speech recognition technology is unavailable, the participant will have to rely on a different service, such as closed captioning (CC), if the event is not in-camera or secret.

Expected results:

  • Increased accessibility to parliamentary proceedings for all participants at debates and committees
  • Increased productivity for interpreters through the availability of an alternative source

Potential challenges:

  • AI systems may produce hallucinations that seem correct.
  • AI systems may not conform to the delays set as standard for CC services by the CRTC.
  • Automated speech recognition systems may be inclined to output content using US or French spelling, or may struggle with specific names or ridings.
  • The automated system favours English output, so may produce results in the wrong language.

Data requirements:

  • Live digital audio feed (in any mode) of the event (public, in-camera and secret)
  • On-premise AI speech recognition systems
  • Dedicated hardware owing to higher processing needs, and ability to transcribe the in-camera proceedings
  • Integrations with other systems:
  • Integration with the live digital audio system
  • Integration with Live Translation App
  • Integration with active directory


Success metrics:

  • Time to access the live transcript text
  • Accuracy of the transcript in each language
  • Ability to recognize the language and produce the transcript in the right language

 

The Use cases for AI in parliaments collection is published by the IPU’s Centre for Innovation in Parliament as part of the Parliamentary Data Science Hub’s project to create guidelines for AI governance in parliaments.

 

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International licence. It may be freely shared and reused with acknowledgement of the author and the IPU. 

 

A use case describes how a system should work. It is used to plan, develop and measure implementation. A use case is not the same as a case study, which is a descriptive text of an actual project’s implementation. Please note that this use case is provided “as is” and neither the IPU nor the author accepts any responsibility for its use.

 

For more information about the IPU’s work on artificial intelligence, please visit www.ipu.org/AI or contact [email protected].