New Orleans Confirms AI Is Now Answering Certain 911 Calls Instead of Human Operators
AI technology is now assisting with non-critical 911 calls in New Orleans, sparking debate over its role in emergency services

New Orleans has confirmed that artificial intelligence is now answering some of its 911 calls. City officials say the technology has been handling repeat emergency traffic reports since 2023.
The admission comes after a post from the More Perfect Union account on X (formerly Twitter) highlighting the system drew more than 32,000 views and hundreds of shares this week. The attention has reignited public debate over AI's growing role in emergency response.
Why 911 Dispatchers Now Have an AI Assistant
The AI does not answer every call to 911. It is triggered only when a caller dials in from within 200 metres of a car accident that has already been logged in the city's dispatch system, according to Orleans Parish Communications District (OPCD) executive director Karl Fasold.
If the AI agent confirms the caller is reporting the same crash, it tells them the incident is already being handled and asks them to stay on the line only if they were involved or have new information. Everyone else is free to hang up.
The district and Carbyne spent three months training the AI on real 911 recordings so it could learn to recognise local speech patterns. Fasold said that teaching the AI to say and understand New Orleans street names, including the notoriously tricky 'Tchoupitoulas', took considerable trial and error before the tool was reliable enough to deploy.
AI is now answering 911 calls in New Orleans.
— More Perfect Union (@MorePerfectUS) August 6, 2026
The AI handles incoming calls and is supposed to analyze and prioritize the calls so that if multiple people phone in about the same incident, the chatbot will handle the repetitive calls about the same incident.
Critics warn that…
Calls Reduced and Dispatchers Freed Up
Fasold said dispatchers were routinely overwhelmed by clusters of calls about the same rush-hour pile-up, which buried more urgent emergencies further back in the queue. The district worked with emergency call software firm Carbyne to build the tool after identifying the pattern in its own data.
Carbyne co-founder Alex Dizengof said the surges followed a clear pattern. 'The pattern was every morning and evening during rush hour, there would be an accident, and hundreds of callers called in about the same thing,' he said.
According to figures from Orleans Parish Communication District (OPCD), for every 20 calls received about a single car accident, the AI now handles between 15 and 18 of them. OPCD spokesperson Jay Vise said the district audited every AI-handled call during the trial phase. 'We audited every call, every interaction, and we're happy to say 100 percent were processed properly,' Vise said. The district did not specify what metrics it used to define 'processed properly'. Fasold added the tool now frees up roughly an extra hour of productivity per dispatcher each day, which he said helps protect staff wellbeing in a high-pressure job.
Tulane Experts Call for Transparency on AI in Emergency Response
Not everyone is convinced the rollout is risk-free. Tulane University professors Aron Culotta and Nick Mattei, who co-direct the university's Center for Community-Engaged Artificial Intelligence, said there is currently no city or national policy specifically governing AI use in emergency response, and that callers are not explicitly told when they are speaking to an AI system.
Mattei said public unease may stem less from the technology itself and more from who is deploying it. 'Squeamishness here might come from the fact that ... your government is doing this, and not something that you bought and brought into your house,' he said, adding that for 'very well-contained applications' the tool could be a useful way to stretch limited public resources.
Professor Rob Lalka, from Tulane University's Freeman School of Business, was blunter about the stakes involved. 'If you miss just one case ... if you get it wrong just once, someone's life is on the line, and that's dangerous,' Lalka said.
Officials insist the AI is strictly limited to non-critical, repeat traffic calls and is never used for violent crimes or life-threatening emergencies. Vise said those calls 'will always be a human being, the first voice you hear'.
New Orleans is among the first US cities to formally confirm AI handling live 911 traffic, and the district is now preparing to expand the same technology to its non-emergency 311 line as well. As more local governments face dispatcher shortages and rising call volumes, other cities are likely to watch closely how New Orleans balances efficiency gains against the transparency and oversight concerns experts have raised. Key questions remain around whether residents should be told outright when they are speaking to a machine rather than a person, and what happens to their call data once it is collected.
© Copyright IBTimes 2025. All rights reserved.

























