A life-saving idea - using AI to help with initial diagnosis
Danish start-up Corti has developed a digital emergency assistant that utilizes deep learning technology. The system analyzes emergency calls in depth to help identify the risk that a patient's heart may have stopped - even while the caller is still on the line.
9 May 2018Share
Corti and Copenhagen Emergency Medical Services (Copenhagen EMS) launched a collaboration around two years ago to combine the experience and specialist know-how of people with the analytical capabilities of artificial intelligence. The ultimate aim was to ensure emergency calls always get the best possible response. To help train the deep learning technology behind Corti AI, Copenhagen EMS rolled out Corti's system under the official emergency services number 112. The AI focuses on identifying cardiac arrests by analyzing the descriptions provided by the caller and even the voice of the caller itself.
Corti provides a digital assistant for emergency services dispatchers that monitors their discussions with callers and helps them pin down important markers in their communication. In doing so, the AI doesn’t just factor in what is being said, but also how it is being said (what linguists refer to as "prosody"), using the caller's tone of voice and breathing to work out how serious the situation is. All the data captured by Corti AI during an emergency call is also archived and compared with emergency calls previously analyzed by the system so that important patterns can be recognized. All these elements come together to help Corti AI assess the situation being reported so that it can support emergency dispatchers with warnings and recommendations.
In fact, the Danish developers are convinced that their AI can diagnose the situation being described better than humans. Reportedly, tests that have yet to be fully evaluated, but comprise more than 160,000 archived emergency calls, have shown that the initial diagnosis of the AI system was right in more than 90 percent of cases. That success rate is around 20 percentage points higher than that of human call handlers. All the same, the developers behind Corti AI aren't looking to downgrade people in the decision-making process, or replace them, but rather to provide the best support possible. Following the pilot phase in Copenhagen, Corti AI is now to be rolled out in four more European cities as part of a collaboration with the European Emergency Number Association (EENA).
Corti ApS (2200 Copenhagen N, Denmark)
Website:
http://www.corti.ai/
Interested in news about exhibitors, top offers and trends in the industry?
Browser Notice
Your web browser is outdated. Update your browser for more security, speed and optimal presentation of this page.
Update Browser