Amazon is building artificial intelligence systems to help run 911 call centers in the United States, a move that could reshape how emergencies are handled. NBC News correspondent Gadi Schwartz visited one such center to see the tools in use and explain how they might support human operators. The effort comes as emergency agencies face high call volumes, staffing gaps, and the need to respond faster with better information.
Amazon is developing 911 call centers operated by AI across the country. NBC News’ Gadi Schwartz tours an emergency call center to see how the technology works and how it will help operators.
Why 911 Is Looking to AI
Emergency call centers handle a wide range of crises, from medical emergencies to fires and crimes in progress. In many places, call takers must collect key facts in seconds and relay them to dispatchers while juggling multiple lines. Leaders say any tool that reduces delay can save lives.
In recent years, many centers have shifted to cloud-based tools and digital maps. The next step has been using speech recognition and automation to assist human staff. The aim is not to replace call takers, but to help them work faster and with fewer errors.
How the System Could Work
In demonstrations seen by Schwartz, AI listens to callers, transcribes speech in real time, and pulls out crucial details like location, type of emergency, and any immediate hazards. The system can surface relevant protocols to the call taker on screen. It may also suggest follow-up questions to close gaps or clarify confusing information.
The human operator remains in charge. They can accept, edit, or reject the AI’s prompts. That oversight is important because callers may be panicked, whispering, or in loud environments. AI can help sort the noise, but people make the final decisions.
Potential Benefits for Operators and Callers
Advocates say this approach could reduce the time between the first ring and the moment help is sent. It could also improve consistency by guiding operators through protocols every time, not just during training. For centers struggling with staffing, AI could ease workloads by handling routine tasks and auto-filling reports.
- Faster triage through live transcription and keyword detection.
- Fewer manual steps as forms complete in the background.
- More accurate locations when combined with mapping and device data.
If successful, the tools might help reduce burnout by cutting repetitive data entry and letting operators focus on the caller’s needs.
Risks and Safeguards
The introduction of AI also raises concerns. Misheard words can lead to errors, especially with accents, multiple languages, or background noise. Privacy and data retention policies must be clear, since recordings and transcripts may be stored in the cloud. There is also the risk of bias if models perform better on certain voices or dialects.
Emergency leaders say guardrails are key. That includes human review, audited decision logs, strict access controls, and tested performance across languages and speech patterns. Many centers also require fail-safes so operations continue if the AI system goes down.
What Dispatch Professionals and Advocates Are Watching
Dispatch professionals want to know how the tools perform during real peak periods, such as severe weather or major events. They are also watching integration with existing computer-aided dispatch systems, which must be reliable and secure. Training will matter: operators need to trust the system without relying on it blindly.
Privacy advocates are pressing for limits on how long data is stored and who can access it. They also want transparent testing results and a path for independent audits. Community groups are asking whether the technology will serve callers who do not speak English or who use nonstandard speech, such as children or people with disabilities.
What Comes Next
Schwartz’s tour signals that pilots are underway and that companies see a market for AI assistance in public safety. Any wider rollout will likely depend on measured results: faster response, fewer errors, and strong public trust. Agencies will also weigh cost, training time, and vendor lock-in.
For now, the approach remains human-led. AI listens, transcribes, and suggests, while trained call takers direct the response. If the tools prove reliable and fair, they could become part of standard 911 operations. If they stumble, agencies may slow adoption until problems are solved.
Amazon’s push puts a spotlight on a critical service that depends on speed and accuracy. The next phase will hinge on transparent testing, clear safeguards, and steady improvements in real-world conditions. Communities should watch how these systems handle language differences, high-stress calls, and data protections. Results from early pilots will show whether AI can help operators send help faster when every second counts.
A seasoned technology executive with a proven record of developing and executing innovative strategies to scale high-growth SaaS platforms and enterprise solutions. As a hands-on CTO and systems architect, he combines technical excellence with visionary leadership to drive organizational success.
