Artificial Intelligence (AI) and Machine Learning (ML)* are increasing in popularity in several industry sectors and becoming available in everyday consumer services, and public safety is no different. They are slowly but surely becoming available for PSAPs to use.

Call for Interest

EENA published a call for interest in the context of an upcoming EENA special project. Interest was high among public authorities, private companies, and researchers. Several priorities for the use of AI and ML were listed, such as language detection and translation, and supporting triage.

Call for applications

After studying the input received in the survey responses, EENA launched a call for applications and invited all those interested to submit  their project idea.

*In the context of this activity, AI is considered the simulation of human intelligence processes by machines and computer systems and ML is considered a branch of AI, which focuses on the use of data and algorithms to continuously improve its accuracy from the experience it gathers from analysing large amounts of data, learning from the insights, and making informed decisions.

Objective

The objective of the project is to explore the use of AI tools in PSAPs and demonstrate their impact, publish lessons learned and recommendations, ensure compliance with EU law, in particular the draft AI Act. The project aims to show how AI tools can be effectively implemented into PSAPs and how they can contribute to increased efficiency and, as a result, a better emergency service for citizens across Europe.

Pilots

The project brings together four companies in the field of AI technology- Gladia, Cestel, Augmented Hearing, LiveReader- and PSAPS from eight European countries, including Finland, Sweden, Germany, North Macedonia, Italy, Spain, Czechia and Portugal. The project has launched ten pilot programmes, with each company working alongside one PSAP, to explore and implement advanced AI solutions in emergency communications across Europe.

 

NumberPSAPCompany
1Finland (SouthWest Finland)Gladia
2Germany (Ludwigshafen)LiveReader
3Italy (Trento)Gladia
4North MacedoniaGladia
5PortugalAugmented Hearing
6.1Spain (SOS Deiak)Cestel
6.2Spain (112 Andalucía)Gladia
7.1SwedenAugmented Hearing
7.2SwedenGladia
8CzechiaCestel

The pilot projects will focus on three critical areas: language detection, translation, and transcription; triage to prioritize emergency calls effectively; and noise cancellation to improve call clarity. These technologies are expected to streamline call handling processes, reduce response times, and ultimately save lives.

More specifically, the pilots will include:

  • Identification, translation, and transcription of emergency calls in real time.
  • Synthesis of translated text in audio.
  • Support the call-taker with hints and recommendations during the emergency call.
  • A speech audio filter for noise reduction and audio filtering of recorded live emergency calls.

Finland

The Southwest Finland Emergency Services will test Gladia’s solution consisting in transcription, language identification and real-time translation of emergency calls.

Germany 

The ALTERNIS solution provided by LiveReader uses several AI technologies to identify the language spoken during a phone call, transcribes the audio to text, translates the text into the listener’s language and synthesises this text into an audio output. The pilot will test this technology in the Integrated Control Centre of Ludwigshagen which handles emergency calls in the Greater Ludwigshafen area. The pilot will consist in simultaneous translation used as an additional service on several work-stations for both real emergency calls and in protected environment with test cases.

Italy

The pilot will take place in the two emergency response centres of the Autonomous Province of Trento. It will consist in using the technology provided by Gladia using artificial intelligence to transcribe live emergency calls received by the centre and asynchronous transcription of the calls in major events.

North Macedonia

The Crisis Management Centre of North Macedonia will use the technology provided by Gladia to transcribe and translate emergency calls received.

Portugal

The emergency calls centre of the Lisbon area will be using Sharpi Box, a technology provided by Augmented Hearing to eliminate background noise during emergency calls, facilitating the work of the call-taker.

Spain

The Emergencias 112 Andalucía centre will use the technologies of Cestel and of Gladia in two separate pilots consisting in automatic translation and transcription to assist the emergency call-taker in different languages. The solution will be incorporated in both the emergency calls management and in testing environment.

Sweden

SOS Alarm, the organisation responsible for handling emergency communications in Sweden will take part in two pilots during this project. In a first pilot, SOS Alarm will test Sharpi Box, a technology provided by Augmented Hearing to eliminate background noise during emergency calls, facilitating the work of the call-taker. The pilot will first focus on testing pre-recorded calls and may eventually proceed to live call situations. In a separate pilot, SOS Alarm will use the technology provided by Gladia consisting in transcription, language identification and real-time translation of calls.

Czechia

The Fire and Rescue Service will use the technology of Cestel for automatic language translation and transcription to help with an increasing number of emergency communications from callers who don’t speak the local language.

Project partners

Companies providing AI solutions

Augmented Hearing is a Danish audio startup enabling new standards for safe spoken communication. Our AI software modules remove noise and clarify speech in real-time during incoming 112 calls, securing a more efficient collection of critical information. Our solutions secure maximum privacy, reliability, and effective AI deployment.

https://augmentedhearing.io

Cestel is one of the Spanish companies with the greatest experience in integrated communication solutions for Emergency, Control and Command Centres, as well as in the video communication channel (videoconferencing, Video Call Centres, Interpretation Centres for deaf people, web video telephony, etc…). In addition to its own products, CESTEL creates unique solutions and specific developments tailored to each client.

https://www.cestel.es/

Gladia leverages advanced AI technology to provide real-time, accurate, and cost-effective audio transcription services. Its AI platform surpasses traditional challenges by offering quick processing, extensive language support, and features like speaker diarization and automatic punctuation, serving a wide array of industries including call centers and media. Gladia’s AI capabilities extend to translation and summarization, enhancing the value of audio data for businesses

https://www.gladia.io

LiveReader GmbH addresses the challenges of knowledge-intensive processes such as the work of dispatchers in emergency call centers with modern AI-driven technologies. ALTERNIS overcomes existing language barriers through simultaneous translation and enables communication in everyday and critical situations. At the same time, the information transmitted by such a phone call is used to control an AI-supported emergency call procedure and provide a holistic solution.

https://livereader.com

PSAPs

Germany – Integrated Control Centre in Ludwigshafen: This PSAP has in recent years dedicated itself to modernizing its control centre processes through modern technologies. The PSAP expressed an interest in solutions which could overcome language barriers, or more generally improve the quality of emergency call processing.

Italy – Provincia Autonoma of Trento: This PSAP manages 1st and 2nd level PSAP (firefighter department and medical department) in the North of Italy. It has 2 emergency responders sites and manages about 300.000 calls every year.

North Macedonia – Crisis Management Center (CMC): The CMC is the national crisis centre of North Macedonia. Within the Crisis Management centre, the State Operation centre functions on a national level through the Emergency Number E-112.

Portugal – Lisbon and Porto PSAPs: These PSAPS are the largest PSAPs in Portugal. Their call takers receive 112 calls and other emergency communications, make the first triage and convey relevant information to the correspondent emergency services.

Spain – Emergencies 112 Andalusia: This PSAP is responsible for all emergency services in Andalusia, and supports callers until the emergency has been resolved or the operatives have ended their interventions.

Spain – SOS Deiak Emergency Coordination Centre: This PSAP is responsible for all emergency services in the Basque Country. With over 100 staff, it receives more than 1 million calls per year, and is also responsible for transmitting public warnings.

Czechia – Fire Rescue Service of the Czech Republic: This PSAP receives all forms of emergency calls, and organises dispatches for all non-medical emergencies. With an average of 36 call taker and 86 dispatchers working there on a daily basis, the Fire Rescue Service responds to over 3 million calls per year, and is also repsonsible for transmitting emergency SMS public warnings.

Sweden – SOS Alarm Sverige AB: This government appointed PSAP isprimarily responsible for emergency and rescue services in Sweden. It receives and forwards urgent 112 calls, and assists municipalities and regions in prioritising ambulance and fire brigade responses.

Finland – South West Finland Rescue Department: This rescue service is responsible for providing emergency services in an area of 20,000 square kilometres on land and water in Finland. Its 600 hundred staff is largely composed of firefighters and emergency medical personnel.

Timeframe

April 2024 (EENA Conference): Official kick-off and press announcement with large press coverage

May – September 2024: Deployments and tests. Short monthly reports by each partner will be requested. One or two web meetings with all partners.

September 2024: Final web meeting

October 2024: Final reporting by each partner, including statistics and feedback on the tools and cooperation, what went well, what went wrong, and lessons learned.

October 2024: Final report on the tests, lessons learned and recommendations. Special note on AI, ethics and data protection.

December 2024: Final public event of the project with 100 – 200 attendees, NG112 & AI Event, 4 – 5 December 2024

Please note, this time frame may evolve or change as required.