Need for Speed – Time Critical In-Language Access to Emergency Management Victoria’s Disaster Warning System
Tracks
Royal Poinciana
| Monday, July 14, 2025 |
| 12:10 PM - 12:40 PM |
Overview
Reegan Key & Tea Dietterich, 2m Language Services
Details
Key Presentation Learnings:
1. Real-time translation is critical for effective emergency response: Time-critical language access ensures that non-English-speaking communities receive and can act on critical emergency information and alerts simultaneously with English-speaking populations.
2. Community trust and verification are key to adoption: Collaboration with communities ensures translation outputs are culturally appropriate and will meet the needs and preferences of the community. User-testing with real-world emergency alerts are key prior to rollout.
3. Scalability and interoperability are critical for rollout: Translation models must be scalable across multiple languages and adaptable to state-specific emergency terms.
Speaker
Tea Dietterich
CEO
2m Language Services
Need for Speed – Time Critical In-language access to Emergency Management Victoria’s Disaster Warning System
Abstract
Emergency Management Victoria’s (EMV) VicEmergency app delivers emergency information to 3.7 million users. However, culturally and linguistically diverse communities are often excluded from direct emergency messaging, leaving them at greater risk and often relying on unofficial community information. Community leaders may distribute information in their own languages and/or adapt messages to make them applicable and relevant to their community. As such, information is usually slow to reach communities and may be incorrectly interpreted.
In 2022, EMV selected 2M Language Services to build an AI-translator to support the development of the multilingual version of the app called VicEmergency Plus. The app was built as a pilot, to test the viability of automated translations of a trained engine and improve accessibility. This presentation will demonstrate learnings from the VicEmergency Plus pilot.
The pilot shows how AI and human linguistic expertise together can support quick action in the event of an emergency. 2M developed a smart workflow and trained a secure and customised Neural Machine Translation (NMT) engine to enable the app to warn multicultural communities in real time and in multiple languages. A human-in-the-loop approach ensures messages are culturally appropriate. The app’s outputs are further built on a glossary that replicate emergency-specific terminology, tone, and language.
Arabic and Chinese (Simplified) were selected as initial testing languages in the development of the app. Linguistic quality assessment reveals that the custom training engine generated 92.5% (Arabic) and 97.5% (Chinese (Simplified)) fluency scores. The app was also evaluated in targeted community workshops and with current VicEmergency users.
As a majority multicultural country, inclusive and accessible information has the potential to save lives. Innovative solutions like the VicEmergency Plus app drive an important change in emergency management to support all communities and is a viable model to roll out across other parts of Australia.
In 2022, EMV selected 2M Language Services to build an AI-translator to support the development of the multilingual version of the app called VicEmergency Plus. The app was built as a pilot, to test the viability of automated translations of a trained engine and improve accessibility. This presentation will demonstrate learnings from the VicEmergency Plus pilot.
The pilot shows how AI and human linguistic expertise together can support quick action in the event of an emergency. 2M developed a smart workflow and trained a secure and customised Neural Machine Translation (NMT) engine to enable the app to warn multicultural communities in real time and in multiple languages. A human-in-the-loop approach ensures messages are culturally appropriate. The app’s outputs are further built on a glossary that replicate emergency-specific terminology, tone, and language.
Arabic and Chinese (Simplified) were selected as initial testing languages in the development of the app. Linguistic quality assessment reveals that the custom training engine generated 92.5% (Arabic) and 97.5% (Chinese (Simplified)) fluency scores. The app was also evaluated in targeted community workshops and with current VicEmergency users.
As a majority multicultural country, inclusive and accessible information has the potential to save lives. Innovative solutions like the VicEmergency Plus app drive an important change in emergency management to support all communities and is a viable model to roll out across other parts of Australia.
Biography
Tea Dietterich is Founder and CEO of 2M Language Services (2M), Australia’s leading language technology and language service provider.
Tea is instrumental in providing secure time-critical language access for Migrant and Deaf & Hard of Hearing Communities in emergency management.
Tea has positioned 2M as a trusted partner supporting the in-language capabilities of Australia’s Warning System to achieve safety equity for non-English-speaking populations.
Tea is President of AALC, Graduate of AICD, has featured in the Australian Financial Review as international Language Technology Innovator, Board and Council Member of University of Queensland BA Advisory Board, Board Member of Austmine, among others.
