Artificial Intelligence in the Emergency Services

Artificial Intelligence is an area of computer science covering the creation of ‘intelligent’ machines, enabling computers to respond to various inputs and even adapt to their surrounding environment via a process called Machine Learning whereby machines identify patterns in data to expand their ‘knowledge’ and ability to perform certain tasks.

The Corti Digital Assistant, developed by Corti & Copenhagen EMS, has already been implemented in Copenhagen, where Corti has been accurate in up to 95% of cases. The EENA has confirmed they will be bringing Corti to four more sites across Europe for testing between September 2018 and April 2019.

Corti works by listening in on emergency calls to analyse information and patterns in verbal and non-verbal signals, such as breathing patterns, tone of voice and verbally communicated information. Corti, in real-time, provides the dispatcher with information, predictions on the criticality of the patient’s condition as well as recommendations and instructions on the correct course of action the patient/bystander should take whilst waiting for assistance, for example CPR Instructions for a patient going into Cardiac Arrest.

Whilst confident in Corti’s ability to make the correct decisions, CTO Lars Maaløe questions its ability to detect nuances or make judgements when introduced to new or complex situations. He gave the example of calling in on behalf of a loved one suffering Cardiac Arrest, the caller would be more confident that the patient is breathing as they want to believe it is true; Would Corti be able to pick up on these human subtleties? Hence why Corti only offers guidance instead of making decisions alone.

The EENA believes that Corti has the potential to save thousands of lives and change the way emergency medical calls are handled and when used in conjunction with dispatchers, significantly reduce error rates and make critical diagnoses faster.

Leave a Reply

Your email address will not be published. Required fields are marked *