🔥 What's In This Issue

• Local wins: Macomb County deploys AI for 911, Boston open-sources data tools, city chatbot reality check, San Bernardino fire sensors.
• States step up: New York trains 100K+ employees, California pilots caseworker AI, states rethink vendor management.
• Best practices: Safe AI pathfinding frameworks, mobile command centers, governance strategies, security awareness.
• Tools to try: Microsoft Copilot expands, ed-tech AI acquisition, Casper bodycam analysis, OpenAI judgment framework.
• Policy watch: Idaho debates DEI chatbot ban, Colorado revises AI law, schools set AI rules, ODNI framework.

🚀 Local Innovators

• AI Will Assist Macomb County, Mich., Dispatch on 911 Calls
  Macomb County approved a $41,000 AI platform that automatically transcribes 911 calls, evaluates dispatcher performance, and identifies trends to improve emergency response training and service delivery.
• Boston's CIO wants the public — and other city governments — to use his open-source AI tools
  Boston CIO Santiago Garces is open-sourcing agentic AI tools that enable anyone to analyze city data through natural language, encouraging other municipalities to adapt the platform for their own open data portals.
• I tested 3 city AI chatbots. Here's what they actually do.
  A hands-on review of AI chatbots from three cities reveals practical limitations and capabilities, showing what works in real-world constituent interactions and where current technology falls short.
• A new kind of fire detection technology comes to a California county
  San Bernardino County deployed lightning strike detection sensors that provide early warnings when high-risk strikes could escalate into wildfires, enabling faster emergency response.
• New York City Public Schools Unveil Guidelines for AI Use
  NYC public schools released comprehensive AI usage guidelines for students and staff, establishing guardrails for responsible classroom integration while encouraging educational innovation.
• CUNY Invests $3M to Support 113 Campus AI Projects
  The City University of New York invested $3 million to support 113 AI projects across its campuses, fostering innovation in teaching, research, and administrative operations.

🏛️ State Innovators

• New York scales up AI tool, training resources to entire state government
  After a successful pilot program, New York State is expanding AI tools and comprehensive training to over 100,000 state employees, representing one of the largest government AI adoption initiatives nationwide.
• AI Lets State and Local Governments Extend Their Reach
  Government agencies are leveraging AI to expand service delivery capacity, reach underserved populations, and provide 24/7 constituent support without proportionally increasing staffing costs.
• Open-source AI assistant shows promise for California caseworkers' service delivery
  California is piloting an open-source AI assistant designed to help caseworkers navigate complex eligibility rules and improve service delivery for vulnerable populations seeking government benefits.
• North Carolina Treasurer Embraces Widespread Use of AI
  North Carolina State Treasurer Dale Folwell is championing AI adoption across the department, using automation and advanced analytics to transform investment management and pension administration.
• Tennessee leans on public-private-sector collaboration — and maybe $50 million — to 'throw a rope' around AI
  Tennessee is investing up to $50 million in public-private partnerships to accelerate AI adoption statewide, with focus areas including workforce training, infrastructure development, and cross-sector collaboration.
• 5 ways state and local governments will operationalize AI in 2026
  Route Fifty identifies five critical strategies for moving AI from experimental pilots to full operational deployment, including embedded automation, grant modernization, and developing AI-proficient staff.

📚 Insights & Best Practices

• Inside DOGE's early days of pressure campaigns, rule breaking and 'chaos'
  Nextgov examines the Department of Government Efficiency's controversial early operations, including aggressive technology modernization tactics and their implications for government AI adoption.
• Safe AI pathfinding is essential for government adoption, officials say
  Government technology leaders emphasize establishing safe, tested AI implementation pathways before scaling adoption, balancing innovation velocity with comprehensive risk management and public trust.
• On the move: How to build an AI-enabled mobile command center
  A practical implementation guide for public safety agencies on integrating AI capabilities into mobile command centers for real-time situational awareness and coordinated emergency response.
• Get Better AI Outcomes Through Information Governance
  Iron Mountain Government Solutions explains why effective information governance—including data quality, lifecycle management, and security—is foundational for agencies deploying AI successfully.
• Government official impersonation scam complaints doubled in 2025, FBI report shows
  An FBI report reveals government impersonation scams doubled last year, with criminals increasingly using AI to create convincing fake identities and communications targeting vulnerable populations.
• After Pilot, NYS Expands Staff AI Training to Over 100,000
  Following successful pilot results, New York State is scaling its AI training program to reach more than 100,000 employees, focusing on responsible use and practical applications.

🦄 Tools, Apps and New Features

• Ed-Tech Nonprofit Acquires Mainstay for AI-Powered Student Support
  A leading education technology nonprofit acquired Mainstay to expand AI-powered student support services, helping colleges improve enrollment, retention, and student success through conversational AI.
• Microsoft expands Copilot agentic tools in government clouds
  Microsoft rolled out expanded Copilot capabilities in government cloud environments, including agentic AI tools that can autonomously execute complex multi-step workflows with proper security controls.
• Casper, Wyo., Will Use AI to Analyze Police Bodycam Footage
  Casper, Wyoming is deploying AI to automatically analyze police body camera footage, helping officers with report writing and evidence review while raising important privacy and oversight questions.
• OpenAI national security lead endorses 'appropriate human judgment' in AI
  OpenAI's national security director emphasized the critical importance of maintaining appropriate human judgment and oversight in AI systems deployed for government and defense applications.
• Most government leaders say they outpace private sector on AI adoption, survey says
  A new survey shows majority of government technology leaders believe their agencies are adopting AI faster than private sector counterparts, driven by mission urgency and constituent needs.
• GSA, NIST partner to craft evaluation standards for AI tools in federal operations
  The General Services Administration and National Institute of Standards and Technology are collaborating to develop comprehensive evaluation standards for AI tools deployed in government operations.

📄 AI Policy & Legislation

• As Las Vegas Police Deploy AI, Privacy Advocates Voice Concern
  Las Vegas police are implementing AI-powered surveillance and analysis tools, prompting privacy advocates to raise concerns about oversight, transparency, and potential civil liberties implications.
• Idaho lawmakers consider banning pro-DEI chatbots inside state government
  Idaho legislators are debating controversial legislation that would prohibit state agencies from using AI chatbots programmed to promote diversity, equity, and inclusion initiatives.
• Sonoma Valley Schools Approve Specific Rules for AI Use
  Sonoma Valley Unified School District adopted comprehensive policies regulating artificial intelligence use among students and staff, establishing clear boundaries and educational guidelines.
• Colorado releases new AI Policy framework aimed revising the state's 2024 law
  Colorado unveiled a revised AI policy framework addressing stakeholder concerns about its 2024 legislation, seeking to balance consumer protection with business innovation and economic competitiveness.
• ODNI is building a framework to boost spy agencies' AI adoption
  The Office of the Director of National Intelligence is developing a comprehensive framework to accelerate AI adoption across intelligence agencies while maintaining security and oversight standards.
• LSU Launches AI Bachelor's Program, 3-Year Degree Track
  Louisiana State University introduced a new bachelor's degree program in artificial intelligence with an accelerated three-year track, addressing workforce demand for AI professionals.

Local Government AI Chatbots: The Reality Check

Every city and county in America is being pitched AI chatbots for 311 and constituent services right now. The vendor demos look great. The promises sound even better.

But a recent hands-on test of three live city chatbots — Denver's "Sunny," Winter Haven's "Ask Winter Haven," and Atlanta's "Ava" — tells a more sobering story.

Denver's chatbot couldn't distinguish a policy question about homelessness from a service request to report an encampment. Winter Haven's couldn't discuss local crime but readily delivered enforcement statistics on homeless camp removals — a political choice, not a technical one. And Atlanta's "Ava," marketed as AI innovation, turned out to be little more than a keyword search engine that returned irrelevant answers to basic questions.

The pattern is consistent: local governments are deploying systems without fully understanding what they've bought, how they've been configured, or what they're saying to residents on the city's behalf. That's a governance failure.

INNOVATION TIP: Before you draft an RFP, define your chatbot's job.

Start by cataloging every topic and high-volume query you want your chatbot to handle — your website index, FAQ, and call/chat logs are great starting points.

Then classify each question as Critical — requiring consistent, accurate, prewritten answers pulled from a vetted corpus — versus Non-Critical, where real-time AI generation is acceptable. Most vendors don't support prewritten answers for critical questions; this distinction alone will eliminate many contenders.

Next, identify questions requiring live API integration with your systems of record. This is another capability that routinely separates capable vendors from the ones who will embarrass you on day one.

Finally, document your minimum technical requirements: hosting environment, citizen login and data security needs, and accessibility/privacy standards such as WCAG 2.2, CJIS, and NIST. Also nail down your expected monthly traffic volume and the channels — website, SMS, voice, social — through which residents will engage.

Good news: you don't have to build your evaluation framework from scratch. The GovAI Coalition's Buyer's Guide to Enterprise GenAI Tools is free resource developed by local government practitioners and is an excellent starting point. The GovAI Coalition also provides templates and questionnaires to streamline your AI procurement.

Use these resources. Your taxpayers will thank you.