Daily artificial intelligence use by US public safety agencies nearly tripled year-over-year, reaching almost 32% of surveyed personnel, according to the sixth annual Verizon Frontline Public Safety Communications Survey. Some 1,035 US public safety professionals took part in the study, which Lexipol Media Group conducted on the operator's behalf.

The figure marks the sharpest single-year shift in the survey's six-year history of tracking technology adoption among police, fire and emergency medical respondents. More than 56% said they expect to use AI daily within three years, framing the technology as a decision-support tool rather than a back-office experiment.

Verizon, understandably, has a commercial interest in that trajectory. The findings arrive from the operator's own public-safety unit, and the survey doubles as a sales argument for the network capacity that AI-assisted dispatch, video analytics and connected robotics demand. The numbers still merit attention, but with the sponsor in view.

What else is climbing?

Drone and robot deployments grew in parallel. Daily use rose from 15% to 26% of respondents — roughly one agency in four — with deployments concentrated in hazardous environments where machines can keep human responders out of danger while feeding back real-time situational awareness.

That near-doubling matters for traffic planning. A drone streaming video into an incident command post consumes far more uplink bandwidth than a voice call, and sustained daily use multiplies the load.

Network resilience anchored the rest of the findings. More than 77% of respondents named resilience and reliability as their main operational requirements — a figure that has barely moved as the technology mix around it has shifted. Some 49% also cited network performance as the biggest factor when choosing a personal device plan.

What did Verizon say?

Kim Mirabella, VP of Verizon Public Sector and Verizon Frontline, framed the results as proof that adoption has outrun expectations.

She said the 2026 findings show technologies including AI and robotics, once considered futuristic, are already essential tools used by first responders every day.

That claim is directionally supported by the usage numbers. What it does not capture is what those tools actually do in the field — the survey reports adoption rates, not outcomes, and offers no data on accuracy, response-time gains or failure modes when AI-assisted systems meet degraded networks.

Why the reliability numbers cut both ways

The survey's most stubborn statistic — resilience and reliability topping operational priorities at 77% — is also its most commercially useful one for Verizon. Public-safety agencies buy network capacity largely on trust, and Verizon Frontline's dedicated infrastructure for first responders is positioned precisely against that requirement.

Yet the same figure constrains the AI story. If more than three-quarters of respondents rank network dependability above everything else, the industry's next task is not more capable models but connectivity that holds when disaster strikes. An AI tool that depends on a backhaul link is only as resilient as the link.

The gap between the 32% already using AI daily and the 56% who expect to do so by decade's midpoint defines the near-term build-out window. Closing it will depend less on model capability, which is advancing quickly, than on whether network reliability — the same priority respondents have named for six consecutive surveys — keeps pace with the traffic these tools generate.