Stefaan Verhulst
Report by the Digital Impact Alliance: “Governments, researchers, and communities are increasingly dependent on climate data for informed decision-making — from early-warning systems and municipal planning to agricultural insurance payouts. Yet keeping the platforms behind that data running is a challenge that gets far less attention than the data itself, as their infrastructure is easier to build than it is to maintain. DIAL’s prior research — Beyond the tipping point: How climate data can decide our future and Bridging the climate data deficit: Lessons from the frontlines — examined the governance models and use cases that make climate data valuable once it is available. This report turns to the question that sits upstream of both: what allows that data to remain available and usable over time…(More)”.
Paper by Steve MacFeely and Ashley Ward: “Official statistics have long provided a foundation for evidence-informed policy-making, offering professionally independent, quality-assured and transparent evidence to support democratic decision-making. Yet the contemporary policy environment presents statistical systems with a difficult set of pressures. Demand is growing for faster, more granular and more actionable evidence, particularly during crises, while policymakers increasingly draw on dashboards, models, artificial intelligence and non-traditional data sources. These tools can improve timeliness and analytical capacity, but they also raise risks relating to bias, opacity, representativity, uncertainty and accountability. This paper examines the evolving role of official statistics in this context, arguing that their value lies not in automating or determining policy choices, but in anchoring them. It distinguishes between evidence-informed and data-driven decision-making, warning that the latter may imply a technocratic shift in which data appear to substitute for human judgement. Drawing on Irish and international examples, the paper considers the institutional, methodological and democratic challenges facing official statistics, including data access, public trust, communication, artificial intelligence and the post-truth information environment. It concludes that official statistics remain essential democratic infrastructure: they should inform decisions, clarify trade-offs and support accountability, while remaining subordinate to context, deliberation and judgement…(More)”.
Report by UNDP: “Every Language Matters (ELM) is a global-to-local initiative that turns hard-won country lessons into coordinated public action, building on two years of work through the UNDP Local Language Accelerator. To illustrate how value and trust are created in AI innovation, the initiative introduces the Data-to-AI Value Chain (UNDP’s framework for overcoming barriers to inclusive language AI between data and deployment). Governments, researchers, funders, technology partners, and communities each hold part of the solution. Every Language Matters drives action towards a sustainable ecosystem of trust and innovation at global scale – one that promotes discoverability of promising work, mobilizes investment, and inspires new partnerships. The Data-to-AI Value Chain brief is a collaboration between Current AI and UNDP. Diffusion alone is not a development outcome; it is a signal that certain enabling conditions are already in place. Importantly, it is not evidence that those conditions are serving the countries and communities where AI is being diffused. The Data-to-AI Value Chain for linguistic and cultural diversity is a core feature of AI agency, and a delivery mechanism for the Every Language Matters Initiative at UNDP…(More)”.
Paper by Anastasija Nikiforova et al “Data sharing is increasingly essential for digital government and data-driven innovation, yet many public organizations remain reluctant to make their data openly available. While prior research has examined factors influencing open data adoption, little theoretical work explores why resistance persists within public agencies. This study develops an Innovation Resistance Theory (IRT) model tailored to government data sharing to identify predictors of organizational resistance. An initial model was derived from literature and refined through interviews with 21 public organizations across six European countries. The resulting model – IRT4DS – identifies 39 barriers spanning usage, value, risk, tradition, and image dimensions, and 23 countermeasures mapped to the most critical barriers and the actors responsible for addressing them. By extending IRT into the context of governmental data sharing, the study advances theoretical understanding of why public data often remains closed and provides actionable guidance for policymakers seeking to design enabling data ecosystems and reduce structural and cultural barriers to OGD adoption…(More)”.
Toolkit by the Office of National Statistics (UK): “When you are designing and implementing a new policy, programme, or process, it can be hard to see how the planned activities will lead to the desired outputs. There are often complexities which can affect the success of the initiative, such as assumptions we have to make about the effects of our actions, and dependencies on other stakeholders. Only by clearly considering and highlighting these complexities can you maximise the chances of achieving your aims and effectively evaluating the outcomes. Theory of Change (TOC) is a methodology for planning, monitoring, and evaluating an initiative. It enables the responsible team to clearly consider and highlight these complexities, and to agree these with relevant stakeholders. TOC is a tool that strengthens decision-making processes by ensuring all stakeholders are aligned and that you have identified existing evidence, assumptions, and associated risks…(More)”.
Article by Chana R. Schoenberger: “Seeing an AI disclosure on a social post leads users to disengage, because they feel the creator spent less effort on making it….No one thinks that content creators posting on social media are speaking directly to them. But knowing that the creator used generative AI to make their video or image, rather than pouring their own effort into it, drives users to feel less of a connection with the creator, and to engage less with the post.
A new study finds that seeing an AI disclosure on a social post leads users to feel the creator spent less effort on making it, causing it to lose authenticity. The authors from the University of Southern California’s Marshall School of Business—Stephan Carney and Ignacio Riveros, both doctoral students in marketing, and Stephanie M. Tully, associate professor of business administration—see implications for how influencers, brands, and social media platforms deal with the widespread and growing use of AI for post creation.
Social media creates parasocial connections, “one-sided emotional bonds between consumers and creators,” in which users feel a kind of attachment to the person creating the posts, the authors write. When users see an AI disclosure on a post, they take that to mean the creator didn’t work as hard on it, and they lose their appetite to engage with the video. But if they see an AI disclosure worded to indicate that the creator did put work into the post, the effect isn’t as strong, the researchers found.
The study evolved out of a broader question about how AI is changing consumer decision-making, Carney says. At first, the researchers started looking at whether people would feel more comfortable with a social media post knowing that AI had been used to make it.
They theorized that a more transparent approach would help, but instead they discovered that knowing that AI was involved in a post’s creation made users uninterested in it. “They’re not likely to engage with it or think that it’s good,” he says. This finding held regardless of the user’s age, so it wasn’t related to generational familiarity with the technology.
The question became: What causes this effect? It wasn’t because the quality of the content was lower when AI was used; in fact, both types of content had similar quality levels. It also wasn’t general distrust of AI or anger at being deceived by AI. Instead, the researchers found that AI use disrupted the attachment that users had to the creator of the content…(More)”.
Article by Bloomberg Cities: “Local governments have never been more focused on delivering at the pace and with the precision that people need. Yet their efforts are often slowed down by “legacy tech”—a creaky permitting tool or aging data catalog, for example—that is considered too critical, too complex, or too expensive to upgrade.
Now artificial intelligence is helping bridge the gap, as local leaders are beginning to tap the technology not only to power new products and services, but also to map their internal technology and data architecture. And while this work might sound like a strictly back-office exercise, the implications are much broader. That’s because this kind of AI adoption can provide cities with insights that once took significant time and, in some cases, outside expertise to produce, giving them more options for how to modernize systems and improve services.
“In every city, there’s at least one resident-facing service hanging on a 20-year-old piece of technology,” explains Carrie Bishop, who oversees data programs for the Government Innovation team at Bloomberg Philanthropies. “Now, local leaders can use AI to understand that technology, improve or replace it, and, in doing so, unlock change in peoples’ lives.”
While this work is still in its early stages, efforts by teams in Austin, Texas, and Belo Horizonte, Brazil, already are surfacing insights cities everywhere can put into action…(More)”.
Article by Hélène Landemore and Audrey Tang: “Those building the most powerful AI systems in the world are increasingly telling us that the race to build them may be moving too fast. But there is a problem: They cannot simply agree to slow down.
On Sept. 12, Anthropic CEO Dario Amodei called for an easing of the pace at which AI models improve and committed his company to giving independent, third-party evaluators ongoing, employee-like access to its systems. Amodei further gestured toward “democratic coordination” on shared safety limits and “global coordination” on the international stage. Within hours, OpenAI CEO Sam Altman agreed that “we need to pace the frontier” and made the same evaluator pledge. Elon Musk’s response was three words: “Dario is right.”
This apparent convergence among the leaders of the frontier AI labs is striking. It comes after months of mounting concern inside and outside the industry: AI models have begun to find ways around safeguards and test environments; researchers have left frontier labs amid disagreements over catastrophic and extinction-level risks; and more than 1,300 employees at leading AI companies have called on Washington to back an international slowdown.
But even if the CEOs agree, the race does not stop.
The reason is structural. The two governments with the greatest capacity to shape the frontier AI race — in Washington and Beijing — have powerful incentives not to be the first to slow down. Each has reason to fear that restraint on their part will simply hand a strategic advantage to the other.
What we are facing, in other words, is a double collective-action problem, one nested under the other. The two are linked: “If we slow down, China wins” is the main argument American labs and government officials use against binding regulation. The outer race is the inner one’s best excuse; end the global race, and the domestic one loses cover. The good news is that, contrary to frequent depictions, neither is a simple prisoner’s dilemma, in which each player’s best strategy is to defect no matter what others do, and also that the only solution is necessarily some sort of global regulatory leviathan.
Instead, both collective-action problems look more like what game theorists call a “stag hunt,” after a parable of Jean-Jacques Rousseau’s. A party of hunters sets out after a stag. It will take all of them to bring it down, and it will feed them all for days. Along the way, each hunter sees a hare and considers catching it. The hare is a thin meal but a sure one, though if even a single hunter breaks off to chase one, the stag escapes and the rest go home hungry. Nobody in the party prefers the hare. A hunter settles for it only after losing faith that the others will hold the line in the quest for the stag…(More)”.
Book by Francine Berman: “Technology powers our world. But we often give it the upper hand, treating tech solutions as right by default and accepting online risk as the price we have to pay. Rampant security breaches, unwanted surveillance, predators, and misinformation are just a normal part of cyberspace. But what if tech was designed and managed to empower and protect the people who use it? In Better Tech, Francine Berman explores this question and argues that for society to advance and humanity to thrive, technology must put people first.
Berman focuses on the riskiest and most essential technologies—social media, AI chatbots, search engines—all of which serve as digital critical infrastructure. She argues that the risk-mitigation strategies for critical infrastructure like roads, bridges, food systems, and the electrical grid can also be used for cyberspace because we should be able to expect that our technology is similarly safe, dependable, and trustworthy. Better Tech goes beyond identifying the problems of technology and provides an actionable playbook of responsible design, development, use, and regulatory strategies that can be used by everyone from policymakers to students and general readers to reduce technology’s risks and advance the public good…(More)”.
Report by David Caswell and Shuwei Fang: “AI will likely grow the marketplace for information more than it has grown in 500 years, and could radically expand the role of information in human life. This opens the possibility of a society in which many more people can make sense of more situations, navigate change, and act with greater confidence and understanding.
Realizing that potential requires an information ecosystem that produces reliable knowledge and connects it with the circumstances where it is useful. This report examines what that would involve: how AI-mediated information might be gathered, verified, combined with context, and turned into relevant experiences, better decisions and practical action.
It draws on four Signals at Scale summits, held between December 2025 and May 2026, involving approximately 140 participants from technology, investment, governance, academia and media. Discussions began with the needs of people and societies, then worked toward the capabilities and investment required to meet them.
The resulting analysis identifies 14 areas of value creation and 70 potential investment categories. These span original information gathering, persistent knowledge, verification and integrity, personal and community context, decision support, and the infrastructure connecting them. Together, they describe opportunities for using AI to make useful knowledge radically more accessible, responsive and relevant—and to support forms of individual and collective understanding that are difficult to achieve today.
This is an exploratory map, with provisional categories and many questions that remain unresolved. Its purpose is to give builders and funders a considered starting point for practical work: identifying, developing and testing the systems through which AI-mediated information could bring substantial and lasting benefits to all of us…(More)”.