Public safety agencies are using AI. Most don't have policies or training for it. New survey data from 1,975 professionals shows where the gaps are.
Article Highlights
AI in public safety is being adopted whether you’re on board or not. Nearly one in four public safety professionals (22.8%) say they have used AI to support operations, according to the 2026 Public Safety Workforce Trends Report from PowerDMS by NEOGOV, which surveyed 1,975 professionals across law enforcement, corrections, emergency communications, and fire/EMS.
The problem is that the workforce is adopting AI practices before receiving critical guidance on how to properly govern AI in public safety.
Half of agencies have no formal AI policy in place. No guidelines. No safety nets. Two-thirds have provided no structured AI training to their workforce. This creates real risk to public safety agencies, leaders, and staff.
Let’s explore what the data shows about AI for public safety, including where agencies are seeing value, what risks emerge when readiness lags, and three concrete steps to put the necessary safeguards in place before diving headfirst into AI implementation.
Most public safety professionals are using AI to manage workload at a time when agencies are already stretched thin. Cutting down on administrative tasks allows staff to put more time into field work and tasks that require more hands-on, interpersonal interaction.
According to PowerDMS’s data, the top three AI applications in public safety are:
|
Use Case |
What It Looks Like in Practice |
|
Report writing and documentation |
AI drafts narratives; public safety professionals review and approve |
|
Data analysis and research |
AI surfaces patterns across large datasets |
|
Internal communication |
AI writes first drafts of emails, memos, etc., to help communicate information |
It’s not a guess that these departments are working overtime to complete necessary admin work. The same survey found 60% of respondents report staffing shortages, and 42% report working 10 or more hours of overtime per week.
AI isn't replacing human work hours. It is being used to relieve the current staff who fill gaps left by understaffing. Instead of working overtime, public safety professionals are using AI to streamline processes and automate tasks.
Survey respondents put it plainly:
Notably, executives and senior leaders report using AI more than frontline roles in public safety, possibly due to a lack of training structure or the convenience of application to assigned tasks. Larger agencies show they are ready to adopt AI in public safety and enjoy the benefits of reduced workload, while smaller agencies are behind on AI adoption overall.
Discover how we help law enforcement by offering training to protect your agency from liability while adopting AI in public safety.
As the report shows, many professionals admit to adopting AI in public safety before they have the guardrails needed to reduce liability risk.
But that doesn’t mean agencies are not concerned about the risks involved. They are well aware that AI can bring on a new set of issues if not used properly:
Awareness of AI risk without governance to address it doesn't reduce exposure. In many cases, it just indicates that staff knew the risk existed and moved forward anyway.
This doesn’t only happen in public safety; it happens across government sectors. Research from the National Institutes of Health found that local officials across government broadly feel underprepared for AI governance as they expand their AI processes.
Nevertheless, just because the readiness gap is a universal issue doesn’t make it any less urgent to have training infrastructure in place.
When AI tools are used without formal policy, training, or oversight, it’s not enough to say you know there are risks. When risks become vulnerabilities in your processes, your agency is ultimately responsible for explaining what went wrong.
Inconsistent usage across the agency. Some officers use AI for report writing. Others have no idea how to use it or what's permitted. There's no standard for which tools are approved and secure, which use cases are appropriate, or how AI outputs should be reviewed before they're used in official documentation. Training can make AI available to all who could benefit.
Defensibility problems in court. AI-assisted documentation without a governing policy becomes a liability. If your agency can't explain how AI was used – and that it was used fairly and within protocol – to produce a report or support a decision, that decision is harder to defend. The output doesn't change. The ability to account for it does.
Data security exposure. Officers using unapproved AI tools may be inputting sensitive information, including CAD data, investigation details, and personnel records, into platforms with no CJIS alignment or data protection guarantees. That exposure may not surface until after a breach or audit. PowerDMS offers a solution for data security and compliance in public safety.
Union and public trust friction. AI use without transparency creates friction with labor organizations and erodes community confidence in your agency's operations. The 2026 survey confirms agencies are aware of these risks. Even still, most haven't built the guardrails to address them.
Public safety agencies are all too familiar with liability avoidance in managing documentation, training compliance, and policy adherence across a workforce. AI adds a new layer of liability risk, but proper governance can help organizations use these tools safely and responsibly.
It’s important that public safety agencies begin addressing the AI readiness gap before it gains even more traction and becomes harder to govern. Here are four areas where agencies can take concrete action now.
Before deploying new AI tools, establish a cross-functional AI committee to guide implementation. This group should include at least one representative from:
By bringing together diverse perspectives, you ensure that AI initiatives are evaluated not just for efficiency, but for security, ethical implications, and impact on personnel. A dedicated committee provides the oversight necessary to vet tools, draft comprehensive policies, and maintain trust across the agency
Having an AI policy in place is the first step. Along the way, you can make necessary changes. At a minimum, it should address:
If you don't have crucial law enforcement policies covering AI yet, start with approved tools and prohibited use cases. Those two elements alone reduce the most common governance failures.
The fact that 66% of agencies from the survey offer no AI training should raise serious concerns about AI readiness and risk management. Training infrastructure should be the first step, and it’s been skipped in most agencies. Training doesn't need to be extensive to be effective, but it does need to reach frontline personnel, not just command staff.
Effective AI training for public safety personnel should cover:
The top-down experimentation pattern visible in the survey means most officers are operating without exposure or guidance. Writing policies and procedures and best practices for law enforcement using AI is vital to learn in any operational area. Training and policy need to reinforce each other.
Some AI tools may be too risky for sensitive information collected in public safety. Focus on adopting AI in areas with clear efficiency gains and manageable risk, such as documentation, training compliance, policy distribution, and data analysis.
When evaluating AI tools, prioritize:
The goal isn't to find the most capable AI tool. It's about finding tools specific to your agency’s needs without creating new compliance or liability exposures.
PowerDMS by NEOGOV uses AI to help public safety leaders manage administrative burden without removing human judgment from the process. Every AI feature is designed with oversight built in and optional use based on your agency’s policies.
Security and governance guardrails built into the platform:
With PowerDMS, you can optimize AI implementation while minimizing risk, knowing that decision-making and oversight stay in the hands of humans. It’s true that AI in public safety reduces documentation burden and strengthens the audit trail – and when done right, you can use AI to enhance agency operations and relieve staff of lengthy admin duties.
Book a demo to see how PowerDMS by NEOGOV can help your agency close the readiness gap and cultivate a protected, resilient workforce.
The most common applications are report writing, data analysis, internal communication, policy writing, and training. These are administrative tasks, not decision-making or surveillance functions.
According to the 2026 PowerDMS Public Safety Workforce Trends Report, 50% of agencies have no formal AI policy in place.
The primary risks are inconsistent usage, defensibility problems in court, data security exposure from unapproved tools, and erosion of public and union trust.
Start by building an AI committee and identifying which AI tools are currently in use, approved or not. From there, develop a policy covering approved tools, permitted use cases, human oversight requirements, and a review cadence.
The data does not support that. Survey respondents consistently describe AI as handling tasks that exceed current staff capacity or consume disproportionate administrative time, not as a replacement for trained personnel.