News That Caught Our Eye #79

Published on October 9, 2025

Summary

This week’s reporting captures a widening gap between AI adoption and the structures meant to guide it, from multilateral arenas to frontline agencies. The World Bank outlines readiness tools for public institutions; the ITU warns of deepening asymmetries in standards, compute, and global voice. National tensions emerge over worker displacement, law enforcement surveillance, and opaque procurement, while researchers and civic actors are experimenting with new participatory frameworks from Mexico to the Basque Country to Northeastern’s “silicon sampling” lab. Together, they sketch a messy yet essential evolution toward accountable, civic-aligned AI.

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.

Coming up this week:

Solving Problems with the Public, October 14, 2:00 PM ET
AI 101 for Communicators, What You Need to Know, Oct 15, 2:00 PM ET
Defining Problems with the Public, October 16, 2:00 PM ET
Smart Data, Fair Cities: Evidence-Based Approaches to City Management, October 16, 2:00 PM ET

You're also invited to join us for an upcoming partner event, "Foundations for the Digital Commons," sponsored by The Roux Institute at Northeastern University in Portland, ME, Oct 30 - 31.

Governing AI

Trustworthy AI Starts with Government Readiness

Jinhee Park, Ahram Han, Chansoo Kim and Haneol Cho on September 2, 2025 in World Bank

A new World Bank report offers a practical roadmap for responsible AI adoption in the public sector, tailored to governments navigating privacy, fairness, and institutional accountability. The paper synthesizes global principles and technical safeguards such as explainability and privacy-enhancing technologies. It equips public institutions with two key tools: a self-assessment flowchart for AI use and a data privacy readiness checklist. Grounded in the realities of state capacity and public trust, the report aims to turn ethical AI from abstract aspiration into operational practice.

Read article

Governing AI

The Annual AI Governance Report 2025: Steering the Future of AI

Staff on September 30, 2025 in ITU/AI for Good

This white paper offers a sweeping synthesis of the year’s major developments in AI governance, including emerging concerns around AI agents, energy use, compute concentration, and global standards coordination. It documents how generative AI is reshaping labor markets, accelerating scientific discovery, and intensifying geopolitical competition. With new attention to verification regimes, risk classification frameworks, and multilateral alignment, the report identifies verification as a cornerstone of credible governance, mirroring approaches in climate and arms control. It also warns that despite growing Track 2 diplomacy, more than 150 countries still lack a meaningful voice in global AI negotiations.

Read article

Governing AI

Our Transparency Rules Need to Adapt to the Rise of AI

Louise Crow on October 7, 2025 in PublicTechnology.net

As the UK government accelerates AI adoption across public services, Louise Crow of mySociety argues that existing transparency mechanisms like Freedom of Information (FOI) are lagging behind. She calls for stronger accountability rules, especially for private contractors, amid rising reliance on opaque, generative systems. Drawing on lessons from failed algorithmic systems and the Post Office Horizon scandal, Crow highlights the power imbalance between public agencies and big tech, and urges legal reforms to extend FOI obligations to AI vendors.

Read article

AI and Labor

The Big Tech Oligarchs’ War Against Workers

Bernie Sanders and Minority Staff on October 6, 2025 in U.S. Senate HELP Committee

A new staff report from the Senate Health, Education, Labor, and Pensions (HELP) Committee warns that AI and automation could displace nearly 100 million U.S. jobs within a decade. Led by Ranking Member Bernie Sanders, the report draws on ChatGPT-based modeling, corporate filings, and labor market analysis to highlight accelerating workforce substitution by “artificial labor.” It criticizes corporate AI deployment strategies and outlines policy reforms, from a 32-hour workweek and robot tax to collective bargaining protections, designed to ensure that workers, not just executives and shareholders, benefit from technological change.

Read article

AI and International Relations

The Next UN: AI, Power, and What Global Governance Must Become

Elana Banin on October 7, 2025 in Reboot Democracy

All 193 UN Member States recently endorsed a global AI resolution, signaling a diplomatic milestone. But will it lead to real progress or reinforce the top-down power structures that have long haunted multilateral governance? Elana Banin, new Editor of the Reboot Democracy Blog, argues that legitimacy must be earned from the ground up. Drawing on examples from California and Vietnam, she outlines three strategic tests the UN must pass to make its AI commitments meaningful: move faster than deployment, fund democratic capacity, and shift power, not just perspectives.

Read article

AI and International Relations

Two Paths to Public AI: Comparing Spain and Mexico’s Strategies

Edgar A. Ruvalcaba-Gomez & Victor H. Garcia-Benitez on October 1, 2025 in Review of Policy Research

A new comparative study analyzes how Spain and Mexico are tackling AI governance in the public sector. Both countries share a strong political intent to promote AI for public benefit, but diverge in outcomes. Spain has built concrete institutions and projects that are beginning to yield policy results, while Mexico’s efforts remain fragmented and under-prioritized. The paper highlights the growing importance of national policy infrastructure in shaping how AI is deployed across sectors.

Read article

AI and Public Safety

Inside Amazon’s Aggressive Push to Get Cops Using AI Surveillance

Thomas Brewster on October 1, 2025 in Forbes

New reporting from Forbes reveals the extent of Amazon Web Services’ involvement in promoting AI-powered surveillance tools to law enforcement across the U.S. and abroad. Through strategic partnerships, AWS has helped market everything from drone surveillance and gun detection to automated police report software and inmate voice analysis. Emails obtained via public records requests show AWS actively advising agencies on grant opportunities and encouraging the adoption of controversial tools like facial recognition and real-time crime center software. The exposé raises pressing questions about procurement transparency, surveillance accountability, and Big Tech’s growing influence over public safety infrastructure.

Read article

AI and Public Engagement

Silicon Sampling: When Communications Practitioners Should (and Shouldn’t) use AI in the Survey Pipeline

John Wihbey & Samantha D’Alonzo on October 7, 2025 in Reboot Democracy Blog

A new post from Northeastern University’s AIMES Lab explores the rise of “silicon sampling,” or the use of LLMs like ChatGPT to simulate public opinion. After reviewing 30+ academic studies, the authors argue that while AI can sharpen early-stage survey design and help identify high-variance questions, it fails as a substitute for real people. Practitioners are urged to take a hybrid approach, treating LLMs as collaborative tools for gathering insight, particularly in politically or culturally sensitive contexts.

Read article

AI and Public Engagement

Democracy in the Digital Era: Digital Technologies and Democracy – An Analysis from the Perspective of Arantzazulab’s Laboratory Activities

Beatriz Belmonte and Diana Pamela Villa on September 30, 2025 in Arantzazulab

A new report from Arantzazulab, a democracy innovation lab in the Basque Country, examines how digital technologies such as AI, Web3, and collective intelligence tools are being used in democratic practice around the world. The publication includes interviews with international experts, a review of more than 40 case studies, and a taxonomy of 66 organizations working at the intersection of technology and democratic participation. It identifies emerging trends, ethical considerations, and practical pathways for aligning digital innovation with democratic governance.

Read article

AI Infrastructure

Feeding the Beast: Powering Democratic AI with Open Data

Beth Simone Noveck on October 6, 2025 in Reboot Democracy Blog

This essay draws a sharp line from the open data policies of the past 15 years to the AI breakthroughs of today and asks what governments should expect in return. With AI hungry for training data, Beth Noveck argues that public datasets are now critical infrastructure and must be treated as such. She makes the case for civic licenses, accountable benchmarks, and investments in the “unglamorous plumbing” of public data. The post features global examples from Abu Dhabi to Helsinki to Indiana, where open data is fueling new services. The analysis issues a call to action for governments to demand reciprocity and build the pipes for democratic AI.

Read article

▲