Seattle Fire Department used Corti AI to triage 911 calls without disclosure
The Seattle Fire Department secretly deployed Corti's AI system to monitor and triage 911 medical calls for over two years without informing callers or conducting public review. The AI prompted dispatchers to route certain callers to a nurse-staffed call center instead of immediately sending ambulances, raising transparency and safety concerns.
What happened
For more than two years, Corti's AI technology has been listening to Seattle Fire Department 911 medical calls without the knowledge of callers. The Seattle Fire Department used the AI to help dispatchers decide which callers should not receive a rapid response, with the technology prompting dispatchers to route certain patients to a nurse-staffed Texas call center rather than send ambulances immediately.
This deployment occurred without disclosure to callers and without public review, confirmed by the Fire Department in late May 2026, two months after The Seattle Times began inquiring about the system. Fire Department officials could not explain how they were measuring the technology's success or provide details about their current contracts with Corti.
Reported impact
Ryan Calo, a University of Washington law professor and co-director of the university's Tech Policy Lab, raised concerns about the use of the company's AI in 911 calls without public knowledge, saying he was 'troubled on a number of levels.'
Index analysis
Experts expressed concerns about transparency, privacy, and accountability in the deployment of AI in medical emergency calls. Callers transferred to the nurse line may wait hours for ambulances, with the city not tracking these delays.
Governance controls
The Seattle Fire Department stated that the dispatcher retains ultimate authority to decide how to handle each 911 call, even when receiving AI prompts, and that every determination is made by trained dispatchers following departmental protocols. Corti said its role is to support dispatcher work rather than replace or override clinical judgment, and stated it does not own the Seattle data and is barred from using it to train models for other customers.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.