How to Advocate for Automatic Speech Recognition in Meetings
Automatic speech recognition (ASR) can make meetings more accessible by converting spoken language into text in real time. When the transcript appears on a screen, participants who are Deaf, hard of hearing, processing spoken language, or joining remotely can follow the discussion, identify speakers, and contribute with greater confidence.
Advocating for ASR is most effective when it is presented as a communication access measure rather than a convenience or replacement for human support. In Australia, meeting practices vary across workplaces, universities, councils, theatres, courts, and community organisations, so a successful proposal should address local expectations, privacy obligations, Australian accents, and the different needs of Deaf and hard-of-hearing people.
| Meeting need | Useful ASR support | Important safeguard |
|---|---|---|
| Following fast discussion | Live captions on a shared screen | Speakers take turns and use microphones |
| Remote participation | Captioned video conferencing | Confirm the platform’s accessibility settings |
| Accurate records | Edited transcript after the meeting | Label it as a draft until reviewed |
| Inclusive decision-making | Searchable keywords and speaker identification | Do not treat automated text as a verbatim legal record |
| Deaf and hard-of-hearing access | Captions alongside interpreting or other adjustments | Ask each participant what access they prefer |
Define The Access Need Clearly
A strong case begins with the people affected. Explain whether participants miss information because several people speak at once, the room has poor acoustics, presenters face away from the audience, or online audio drops out. Someone may hear sound but still need captions to understand unfamiliar names, technical terms, or rapid discussion.
Use specific examples instead of general claims about inclusion. A staff member might need to read a question before responding, while a Deaf participant may use Auslan and prefer an interpreter, captions, or both. ASR should be described as one part of an individual communication plan. It cannot assume that every Deaf person has the same language, identity, or access preference.
In Australia, it is useful to mention the difference between Auslan and English. Automated English captions may assist many people, but they do not translate Auslan. Where an Auslan interpreter is required, live captions should complement interpreting rather than displace it.
Build A Practical Business Case
Decision-makers are more likely to approve ASR when the proposal links access with ordinary meeting outcomes. Captions can reduce requests for repetition, help remote attendees keep track of complex discussion, and make action items easier to locate. They can also support people who are neurodivergent, learning English, or working in a noisy environment.
Set out the cost, equipment, staffing, and expected result. A small organisation might begin with a captioning feature already included in its video platform. A large employer may need dedicated microphones, a caption editor, or a trained captioning provider. If a committee requires a well-structured, referenced briefing, academic writing assistance may help with presentation and research, but the access proposal should still be based on consultation with the people who will use it.
A pilot can make the request easier to approve. Choose several meetings, record whether captions were available, gather feedback, and compare errors across different speakers and room conditions. Report practical measures such as participation, missed agenda items, correction time, and user confidence.
Explain Accuracy And Human Oversight
ASR systems are statistical tools, not neutral observers. They may misrecognise names, Aboriginal and Torres Strait Islander place names, medical terminology, legal language, or words spoken with an accent. Background noise, overlapping voices, poor microphones, and a speaker who moves away from the microphone can quickly reduce accuracy.
Australian English also has regional pronunciation and vocabulary that may not perform consistently across software products. A meeting in Melbourne, Perth, or Cairns can include local references that automated systems have not learned well. Test the chosen platform with the actual meeting room, microphones, participants, and subject matter before relying on it for an important decision.
Human oversight should be part of the proposal. A trained operator can correct key terms, mark inaudible sections, identify speakers, and alert the chair when the transcript is becoming unreliable. For a legal, medical, disciplinary, or high-stakes meeting, an automated transcript should never be presented as a certified or complete record without appropriate review.
Address Privacy And Legal Responsibilities
Before enabling captions, identify what happens to the spoken data. Ask whether audio and transcripts are stored, where servers are located, who can access them, and how long files remain available. Meeting organisers should explain these practices clearly and obtain consent where required, particularly when sensitive health, employment, legal, or personal information may be discussed.
A privacy statement should cover live captions, downloaded transcripts, meeting recordings, and any third-party software. Participants should know whether the transcript will enter the official record and how they can request corrections. Confidential meetings may need an approved enterprise platform, restricted access, or a human captioning service with a clear confidentiality agreement.
Accessibility can intersect with legal duties under Australian discrimination law and workplace obligations. A request for captions should be considered as a reasonable adjustment rather than dismissed because the technology is imperfect. For meetings involving complaints, contracts, or formal proceedings, legal interpreting guidance can help organisations distinguish between general communication access and the specialised requirements of legal interpreting.
Establish Meeting Rules That Make Captions Work
Technology performs better when people change their meeting habits. The chair should ask participants to speak one at a time, use the microphone, state their name when joining the discussion, and avoid talking while another person is presenting. Slides, acronyms, names, and specialist vocabulary can be provided in advance so a captioner or system can prepare a custom dictionary.
The person using captions should be able to see them without turning away from the speaker or losing access to shared documents. In a hybrid meeting, captions may need to appear on a separate monitor or in a dedicated window. The facilitator should pause after a question, allow time for reading, and avoid making decisions while an access delay is still being resolved.
These practices are relevant in a Sydney boardroom, a Brisbane community centre, or a local council meeting in regional New South Wales. They also reflect the direct, practical communication style common in Australian workplaces: explain what is needed, agree on a process, and follow it consistently. A short briefing for staff can prevent captions from being treated as an afterthought.
Include Deaf People In The Evaluation
The best advocacy includes Deaf and hard-of-hearing people from the beginning. Invite them to test the platform, review the meeting layout, and describe what makes the captions useful or frustrating. Consultation should not be limited to whether words appear on the screen; it should consider turn-taking, speaker identification, response time, lighting, interpreter placement, and the emotional effort involved in correcting repeated errors.
Some participants may prefer Auslan interpreting, some may prefer captions, and others may use both depending on the setting. A person who communicates through Auslan may find English captions helpful for written material but insufficient as a substitute for their first language. Captions should therefore sit within a broader accessibility policy that includes interpreters, hearing loops, remote participation, accessible agendas, and plain-language documents.
ASR can also be valuable during unexpected situations. Clear visual information matters when a meeting moves quickly because of an evacuation, medical incident, or public safety concern. Guidance on communication in emergencies can help staff understand why visual cues, gestures, writing, and calm attention remain important when automated systems are unavailable or unreliable.
A final evaluation should record what worked, what failed, and what changes are needed. The aim is not to prove that ASR is perfect. It is to create a dependable access pathway in which technology, skilled people, respectful meeting behaviour, and the preferences of Deaf and hard-of-hearing participants work together.