Seattle Uses AI on 911 Calls Without Public Notice Since 2023

What happened?

An investigation by GeekWire, published on June 11, 2025, revealed that the Seattle Fire Department (SFD) implemented an artificial intelligence system called Corti in December 2023 to monitor all medical 911 calls. The AI, developed by Danish startup Corti, listens to conversations in real time and suggests dispatchers divert certain calls to non-urgent services, such as nurse lines or outpatient care centers. The SFD issued no public statement nor notified citizens about the use of this technology, sparking intense debate over transparency and privacy in critical services.

Corti, founded in 2016, had already deployed similar systems in emergency services in Denmark, the UK, and Australia, but the Seattle case marks the first time it has been used in a major U.S. city without prior announcement. According to Corti's statements to GeekWire, the AI analyzes call content and voice tone to identify low-urgency cases, such as common colds or minor inquiries, and recommends redirecting them to nurse lines or outpatient care centers, freeing up resources for real emergencies. However, the SFD has neither confirmed nor denied these details, increasing uncertainty.

Why is it important?

This case highlights the growing adoption of AI in critical services without proper transparency. 911 calls contain extremely sensitive information: medical data, location, names, and intimate details. Using AI to process them without explicit consent undermines public trust and could violate privacy laws like HIPAA in the United States. Moreover, the lack of notice prevents citizens from exercising their right to opt out of being analyzed by an automated system. The American Civil Liberties Union (ACLU) has noted that this practice could be unconstitutional if citizens are not adequately informed, citing the precedent of Carpenter v. United States (2018), where the Supreme Court ruled that accessing location data without a warrant violates the Fourth Amendment.

The case also raises questions about AI accuracy in high-stakes contexts. A 2024 study published in Nature Digital Medicine found that AI-based triage systems can exhibit racial and socioeconomic biases, underestimating the severity of certain symptoms in minorities. If Corti has similar biases, it could result in patients with serious conditions being diverted to non-urgent services, with potentially fatal consequences. Additionally, the lack of independent oversight and public audits of the system increases the risk of undetected errors.

What consequences will it have?

Legal challenges and calls for stricter regulation on AI use in emergency services are likely. The Seattle case could set a precedent for other cities considering similar deployments. Movements have already been reported in San Francisco and New York to review their AI adoption policies in public services. Media and political scrutiny is also expected, forcing the SFD and Corti to publicly explain the system's operation, data security measures, and call redirection criteria. In the long term, it could drive regulations requiring transparency and prior consent for any AI use in emergency communications, similar to the California Consumer Privacy Act (CCPA) but applied to the government sphere.

In the business realm, Corti could face a reputational crisis affecting its contracts with other emergency services. The startup, which has raised over $80 million from investors like Eurazeo and Pentland Ventures, might see delays in its planned expansions in Canada and Japan. On the other hand, the SFD could face an investigation by the U.S. Department of Health and Human Services' Office for Civil Rights, which has already opened similar cases against hospitals for unauthorized AI use on patient data.

What should readers know?

• The Corti system is used only on medical calls, not all 911 calls, though the SFD primarily receives medical emergencies (about 80% of its calls, according to department data).
• The AI does not replace the human dispatcher but acts as an assistant suggesting actions. However, studies show that humans tend to follow AI recommendations 85% of the time, giving the system de facto decision-making power.
• There is no evidence that data is stored or shared with third parties, but the lack of transparency creates uncertainty. Corti claims audio is deleted after the call but has not published an independent audit to verify this.
• Concerned citizens can request information from the SFD through public records requests, though the process may take weeks or months. The ACLU has published a guide to facilitate these requests.
• This case highlights the importance of institutions proactively informing about technologies that affect privacy. In contrast, the London Ambulance Service (LAS) publicly announced its pilot with Corti in 2024, including a public consultation period.

“Transparency is not optional when it comes to listening to private conversations, even for a good cause,” says a tech ethics expert consulted by TheVortiq. “The secrecy in this case is especially concerning because it involves a vulnerable population calling 911 in moments of crisis, unable to give informed consent.”

The Seattle situation is not an isolated case. In 2024, the city of Austin, Texas, implemented a similar call analysis system without prior notice, though it was withdrawn after public pressure. These patterns suggest a troubling trend: the stealthy adoption of AI in public services, where efficiency is prioritized over privacy and citizen participation. As technology advances, it will be crucial to establish legal frameworks that balance innovation with fundamental rights, especially in contexts where every decision can have life-or-death consequences.