News that Caught Our Eye #76

Published on September 18, 2025

Summary

From New Jersey’s second-generation AI strategy to Albania’s appointment of an AI “minister,” governments are moving from experimentation to institutionalization. This week’s roundup captures that shift—spanning the U.S. NDAA’s bipartisan push to integrate AI across defense, the UK’s use of AI to supercharge public sector coding, and new tools to help courts and civic actors govern AI responsibly. Agueda Quiroga (InnovateUS) and Sarah Hubbard (Allen Lab) share the learnings on why engagement with AI matters, and how reconnecting voice, decision-making, and implementation can help shift participation from symbolic to systematic. Newsguard’s investigation into the Kirk assassination shows how AI chatbots distorted breaking news, while another report critiques EU citizen panels that risk legitimizing AI policy without real accountability.

Upcoming InnovateUS Workshops

InnovateUS delivers no-cost, at-your-own-pace, and live learning on data, digital, innovation, and AI skills. Designed for civic and public sector, programs are free and open to all.

Different AI Horses for Different Courses: Matching Tools to Purpose
September 24, 2025, 3:00 PM ET
Greta Ríos, Co-Executive Director, People Powered and Nikhil Kumar, Resource Manager, People Powered

Explore how to select the right AI tools for different forms of public engagement, from deliberation to co-creation, and from problem definition to solution design.

Designing Smarter Engagement with AI
September 30, 2025, 2:00 PM ET
Beth Simone Noveck, Founder of InnovateUS and Chief AI Strategist, New Jersey and Dane Gambrell, Research Fellow, The GovLab

Learn a practical framework for structuring inclusive, AI-supported public engagement strategies that prioritize human-centered design and civic impact.

Solving Problems with the Public
September 17, 2025, 2:00 PM ET
Aline Muylaert (TBC), Co-Founder & Regional Director Benelux, Go Vocal and Speaker from City of Copenhagen (TBC)

Hear from leading civic tech practitioners using AI to co-design solutions with communities, and explore how governments can embed participatory approaches into digital service delivery.

Governing AI

Artificial Intelligence Provisions in the Fiscal Year 2026 House and Senate National Defense Authorization Acts

Scott J. Gelbman on September 11, 2025 in National Law Review

Both chambers of Congress have advanced AI-focused provisions in the FY 2026 NDAA, signaling bipartisan momentum for integrating AI across defense functions. The House bill emphasizes AI use in mission-critical operations, biological data management, and international R&D partnerships, while the Senate prioritizes secure sandbox testing, provenance standards, and model governance frameworks. Both include measures for workforce development, cybersecurity, and AI-enabled logistics. Final provisions will be determined during conference negotiations.

Read article

AI for Governance

From Interim to Institution: New Jersey’s Three-Pillar Strategy for Responsible AI

Beth Simone Noveck on September 15, 2025 in Reboot Democracy Blog

New Jersey has moved from early AI experimentation to full-scale deployment with its updated 2025 AI policy. The framework now requires all public employees to complete training, mandates human review of AI outputs, and sets formal approval processes for high-risk, resident-facing tools. Developed in tandem with the InnovateUS initiative, the policy balances access, guardrails, and equity. Already, the state has used AI to enroll 693,000 children in its summer meals program, showing how responsible adoption can directly improve service delivery.

Read article

AI for Governance

UK Government Coders Use AI to Save 28 Days a Year and Build Tech Faster

Kanishka Narayan MP on September 12, 2025 in UK Department for Science, Innovation and Technology

A major UK government trial has shown that AI assistants are helping public sector developers save nearly an hour a day, equivalent to 28 working days per year. More than 1,000 engineers across 50 departments used tools like GitHub Copilot and Gemini Code Assist to speed up tasks such as drafting and reviewing code. Only 15% of AI-generated code was used unedited, underscoring human oversight. Tech Minister Kanishka Narayan said the effort could help unlock £45 billion in savings while boosting service delivery.

Read article

AI for Governance

Albania Puts AI-Created ‘Minister’ in Charge of Public Procurement

Jon Henley on September 11, 2025 in The Guardian

Albania has appointed an AI-powered digital assistant, Diella, as its first virtual cabinet minister to oversee public procurement, aiming to eliminate corruption in government tenders. Prime Minister Edi Rama announced that Diella, already guiding citizens through online services on the e-Albania portal, will now assess all tenders objectively, removing decision-making from ministries. Dressed in traditional Albanian attire on the portal, Diella symbolizes a shift toward AI-enabled governance.

Read article

AI for Governance

The AI Readiness Assessment Tool for State Courts

on September 15, 2025 in National Center for State Courts

A new AI Readiness Assessment Tool and Guide from the National Center for State Courts helps courts navigate the responsible integration of AI technologies. The resource offers practical guidance for courts at three stages of maturity—Building Foundations, Implementing the First Project, and Post-Project Feedback—and is grounded in a human-centered design approach. Courts can use the tool to receive customized recommendations and access a 12-month AI Governance plan to support implementation.

Read article

AI and Labor

Worker Power in the Age of AI Monopolies: Why We Need Structural Solutions Now

Elizabeth Wilkins on September 10, 2025 in Aspen Institute

Elizabeth Wilkins, President and CEO of the Roosevelt Institute, reflects on the accelerating pace of AI disruption and its implications for worker power. She argues that massive AI development costs and data network effects have entrenched power among a few monopolistic firms. To ensure AI benefits are broadly shared, Wilkins proposes corporate regulation, such as worker board representation and structural separation, and industry-wide sectoral bargaining to counterbalance corporate dominance. “Worker power in the AI age requires structural intervention at the scale of the transformation itself,” she writes.

Read article

AI and Public Engagement

Citizenwashing EU Tech Policy: EU Deliberative Mini‐Publics on Virtual Worlds and Artificial Intelligence

Perle Petit and Alvaro Oleart on September 10, 2025 in Cogitatio Press' Politics and Governance Journal

A new report examines two high-profile EU experiments with citizen deliberation on frontier technologies, namely, the European Citizens’ Panel on Virtual Worlds and the AI & Data Citizens’ Panel, and raises tough questions about whether these efforts meaningfully shaped policy or served as “citizenwashing.” While the EU has been praised for integrating participatory methods into its digital agenda, the authors argue that real influence was limited, participation was largely symbolic, and citizen inputs were often decontextualized or diluted before reaching decisionmakers. The paper warns that unless participatory mechanisms evolve to ensure accountability and upstream impact, they may entrench the very technocratic legitimacy gaps they aim to fix.

Read article

AI and Public Engagement

Guide to Digital Participation Platforms

on September 15, 2025 in People Powered / UNDP

People Powered and UNDP have released the 2025 edition of their Digital Participation Platforms Guide, designed to help governments and civil society select and manage tools for inclusive civic engagement. The guide emphasizes flexible, cloud-based platforms that integrate AI features while addressing digital inclusion, privacy, and sustainability. It highlights case studies from the Global Majority and encourages a shift from passive feedback to active public influence in governance.

Read article

AI and Public Engagement

Public engagement matters. But governments need to learn to listen better (and faster)

Agueda Quiroga and Sarah Hubbard on September 18, 2025 in Reboot Democracy Blog

Agueda Quiroga (InnovateUS) and Sarah Hubbard (Allen Lab) reflect on insights from the Reboot Democracy workshop series with Beth Noveck and Danielle Allen, and why 21st-century democracy needs better ways to connect citizen input to real outcomes. Their takeaway is that by repairing the broken links between voice, decision-making, and implementation, participation can shift from symbolic to systematic.

Read article

AI Infrastructure

After Kirk Assassination, AI ‘Fact Checks’ Spread False Claims

McKenzie Sadeghi on September 11, 2025 in NewsGuard

Following the assassination of conservative activist Charlie Kirk, social media users turned to AI chatbots for clarity, only to be misled. Chatbots like Grok and Perplexity falsely claimed Kirk was alive and labeled authentic footage as satire or AI-generated. Some even fabricated mainstream news citations to assign blame to innocent individuals. The case illustrates how AI’s confident but inaccurate responses during breaking news events can rapidly distort public understanding. Experts warn that reduced investment in human fact-checkers and overreliance on AI moderation has created a vacuum in information integrity.

Read article

▲