By Police1 Staff
DENVER — Police leaders were cautioned to take a deliberate, informed approach to purchasing technology that incorporates artificial intelligence during the “Buyer Beware — Guidance for the Assessment, Procurement and Contracting of AI” session at the International Association of Chiefs of Police Conference.
Panelists emphasized that while AI can improve efficiency and effectiveness, it also introduces new risks related to transparency, bias and legal compliance. Most AI tools used in law enforcement are developed by third-party vendors, often with limited visibility into how they function. That lack of transparency, presenters said, conflicts with the trust and accountability central to policing.
| RELATED: AI in public safety: 10 key questions to ask your technology partner (checklist)
Procurement and policy challenges
Legal advisor Don Zoufal of CrowZnest Consulting urged agencies to integrate AI considerations into all new technology contracts.
“You’re not buying AI,” he said. “You’re buying a product that has AI in it or will have AI in the future.”
Zoufal compared traditional and generative AI, describing the latter as “an innovative chef that creates something new,” but noted that generative systems remain largely untested and poorly understood. Agencies, he said, must account for AI’s evolving nature and ethical implications when evaluating products.
The presentation’s learning objectives focused on three agency priorities: evaluating current technology portfolios for AI use, developing policy and contract practices tailored to emerging technologies, and managing third-party vendors that supply AI-enabled tools.
Tools and guidance available
Presenters cited several national resources to guide agencies, including:
- The IACPImplementing Technology policy guide and Technology Policy Framework;
- NIST’s Artificial Intelligence Risk Management Framework (AI RMF 100-1) and its 2024 Generative AI Profile (AI 600-1);
- The DHS Playbook for Public Sector Generative AI (January 2025); and
- Recent White House and Office of Management and Budget memoranda —M-25-21andM-25-22 —outlining acquisition lifecycle stages, testing and evaluation requirements, and provisions for ongoing vendor monitoring.
However, panelists agreed that overarching federal regulation is lacking. “The regulatory environment from the federal level is going to continue to be absent,” one noted. Instead, agencies must look to state-level developments. In 2024, 45 states and territories introduced AI-related bills, with at least 10 enacting laws to assess or regulate government AI use.
Smarter hires, instant reporting and more — learn how AI is changing the face of law enforcement in the video below.
Case study: Chicago Police Department
A case study from the Chicago Police Department illustrated AI’s potential benefits in training. The agency is exploring AI-enhanced scenario-based instruction using virtual and augmented reality. AI could make simulations more dynamic by adjusting role-player behavior in real time and evaluating officer performance — a cost-effective way to improve realism and feedback.
Texas DPS perspective
Jessica Ballew, CIO, and Michael Parks, Director of Procurement and Contract Services for the Texas Department of Public Safety, shared their experience purchasing software that includes AI capabilities. They said AI procurement “tests the limit of our traditional procurement framework.”
Texas DPS maintains an inventory of more than 50 AI-enabled tools. Each new system is evaluated against policy and risk, using structured pre-procurement and vendor review tools. The agency asks questions such as:
- How will the tool be used and scaled?
- What data will it access, and how is that data maintained?
- Who will use the tool, and what training will they receive?
- Has the AI been tested for bias, and what risks or limitations are known?
Ballew also stressed the importance of “keeping a human in the loop” and consulting other agencies for lessons learned beforecontracting fornew technology.
| ON-DEMAND WEBINAR: Harnessing AI in law enforcement: Transforming digital investigations for smarter policing
Practical tools for agencies
The presentation outlined a practical toolkit for agencies beginning or expanding AI use, including contract risk assessments, AI inventories, data sheets, architecture review boards and AI committees. These tools help structure oversight, ensure vendor accountability and document the lifecycle of AI adoption.
Questions to ask before buying an AI product
- What problem are we trying to solve, and why is AI the right solution?
- What data will the tool access, process or generate — and who controls that data?
- Has the AI system been tested for bias or accuracy, and can the vendor provide results?
- How will the system’s performance be monitored, updated and audited over time?
- Who will use the tool, and what training will they need — on both the product and policy?
- Does the system require human oversight (“human in the loop”) for critical decisions?
- What happens if the system fails or produces an incorrect result — who is accountable?
Bottom line
AI procurement challenges traditional frameworks and requires new approaches to risk, policy and oversight. Agencies must invest in education, policy development, and structured evaluation tools to responsibly adopt AI while preserving public trust.
Police1 is using generative AI to create some content that is edited and fact-checked by our editors.
