Stefaan Verhulst

Paper by Vera Lummis, Angela Aristidou and Rob Reich: “Frontier AI labs and nonprofit organizations are increasingly partnering to deploy frontier AI in support of nonprofits’ missions to improve global health, climate, humanitarian, and development outcomes. Individual and institutional funders have invested hundreds of millions of dollars to support nonprofits in deploying AI for good, occasionally in formal partnership with AI labs. However, the rapidly growing AI for Social Good (AI4SG) ecosystem lacks a clear understanding of what these partnerships provide, how they form, and what drives or compromises their success. 

Drawing on interviews with 36 individuals across 24 organizations (frontier AI labs, nonprofits, philanthropic funders, and independent experts) and four in-depth case studies, this report offers insights into how nonprofits and frontier AI labs partner to deploy AI to produce social impact and provides recommendations for how partnership and funding arrangements can be structured for success…(More)”.

Paper by Victoria Gorton a and Morgan Currie: “This paper outlines benefit recipients’ views towards the sharing of data that is collected about them when they routinely interact with social security services. Sharing this administrative data between public bodies promises to transform the delivery and evaluation of social security. However, there remains a lack of research into public attitudes, understandings, and aspirations for the sharing of this data, particularly from those whose data would be shared.
Drawing on discussions from deliberative focus groups with 41 individuals, we reflect on the scope for increased data sharing to deliver improvements for recipients. While participants were aware of how their data is currently shared and how increased data sharing might have advantages, they expressed strong distrust in government agencies to manage their data and concerns about being placed in a position of increased vulnerability. We argue individuals’ views towards data sharing cannot be separated from their lived experiences of data systems – their data realities – which are often markedly different from system designers. We conclude by offering considerations against data systems re-introducing paternalistic
policy, unintentionally, via automation – which we term ‘automated paternalism’…(More)”.

Article by sheets.works: “Billions of phones run on code looked after by a handful of people. We counted them from the code itself…

People post this comic whenever something breaks. The tower is labelled “all modern digital infrastructure” and it stands on one small block, “a project some random person in Nebraska has been thanklessly maintaining since 2003”.

We wanted to know how close to true it is. We downloaded the full history of 23 pieces of software that phones, browsers and servers depend on, and counted everyone who made ten or more changes to each one between October 2025 and October 2026. Then we listed what is actually installed on an Android phone, an iPhone and a Windows PC…(More)”.

Book by David Weinberger: “Amid the clamor about AI, Beautiful Particulars poses a unique question: Not how does our use of AI change what we do, but how is the idea of AI already subtly yet fundamentally reshaping how we think about ourselves and our world?

In clear, engaging, and even humorous prose, David Weinberger offers a new perspective on how machine learning works and how it may already be changing our basic and deepest ideas. Through lively examples, he shows that AI’s success—with all its risks and dangers—is taking us past our millennia of reliance on laws and principles by revealing what morality, free will, creativity, causality knowledge, and even reality itself look like now that we’re able to let the particulars speak their tiny truths.

The author poses questions not usually asked: What does a coin look like to AI? When AI gives you evidence that the earth is round, why are its own reasons for saying that so different? Why does an AI image generator create a “selfie” of ancient Aztec warriors with broad smiles that it “believes” are historically unlikely? If AI becomes conscious but bodiless, could it still make good moral decisions? Why is the world so much better at being causal than AI could ever “understand”? How does AI show us that genuine creativity is not about a “spark” but about moments that feel right? And how can you take the temperature of a joke?

Beautiful Particulars offers a new—and hopeful—perspective on how machine learning works, and how that’s already changing how we understand what it means to be human in the age of AI…(More)”

Article by Joseph Foti: “In the prior OGP Horizons piece in this series, we looked at what governments actually mean when they talk about tech sovereignty. In the abstract, this is an ability to control technology in their borders and to be relatively autonomous when it comes to using foreign technology.

But, in practice, governments are quite diverse. As a recap, we looked at the legislation and policies being adopted by many “middle powers,” as well as the EU, the US, and China. (Readers are encouraged to skim the summary of all of the findings here.) The findings showed that interventions fell into three categories.

• Cost: Buying software and computing power and spending money to build domestic capacity.
• Exit: Preserving the ability to leave a bad relationship.
• Leverage: Controlling holdings that other jurisdictions need.

Within each of these categories, there are a number of different policies. In Part II of the series, we will go through cost, exit, and leverage. We’ll also look at why and where open government can help ensure that people are getting what they need from the policies, namely autonomy, good value, security, and jobs.

For each of the given policy areas, there is a distinct role for transparency and public oversight. In some cases, especially to those related to land and permitting as well as standards development, there is a particular role for public participation…(More)”.

Book by Catarina Zuzarte Tully: “This book is different.It is written for people who are actively trying to transform the systems around them in a time of uncertainty and change. It is written for people who are experiencing the present as a period of frustration, anxiety and retrenchment—but who also know that it is a time of extraordinary possibility to shape a positive future. It is written for people who understand that transformation is ultimately about power: speaking truth to it, challenging it and redistributing it. It is written for people who feel a sense of stewardship: a responsibility to connect people working across past, present and future in service of a more prosperous, secure and sustainable world.

Future generations are not strangers who happen to live later. They are extensions of ourselves, just as we are extensions of those who came before us. Playing with Time is rooted in this simple idea. It invites us to loosen our grip on the present, reconnect with longer-term horizons, and allow possible futures to reshape how we act today.

This moment calls for our creative, collective response. It asks us to articulate and defend the values we care about, work across difference wherever possible, and act for the wellbeing of current and future generations. These are things that all of us can do. This book shows how...(More)”.

Report by the Wellcome Trust: “The Wellcome Global Monitor is the world’s largest study of how people think and feel about science and health. The 2026 edition draws on surveys of more than 140,000 people in 140 countries, making it a uniquely rich and globally comparable data set on public trust in science. 

This is the third wave of the Wellcome Global Monitor, following surveys in 2018 and 2020. Over the course of these surveys, we saw trust rise during the pandemic, when science was more visible in people’s lives, but that rise has not been sustained. Today, trust in science is below 2018 levels globally. 

Global averages only tell part of the story. Trust levels vary widely between and within countries, and it is closely linked to people’s everyday experiences of healthcare, of public institutions and of whether they feel they are getting by. 

The full report, the dataset, country and regional factsheets and an interactive dashboard are all freely available…(More)”.

OECD Report: “Governments are under growing pressure to respond more effectively to rapid technological, economic and societal changes while maintaining stability and fiscal sustainability. A significant transformation of the functioning of governments needs to take place that can take advantage of digital and AI tools, and the opportunity of large cohorts of retiring civil servants to change the culture. Based on the analysis of cutting-edge practices in governments of OECD countries, Government Unstuck provides a coherent set of actions that will help governments be up to the challenges of today…(More)”.

Book by Gary Wolf: “Science has given us tools and methods that allow us to observe things that were once imperceptible—and this book shows what we can learn when we use them to look at ourselves.

Written by science journalist Gary Wolf, co-founder of the Quantified Self movement, this book teaches readers how to use proven scientific methods to answer their own health questions. At a time when trust in expertise is fractured and people are divided not just politically but over how to know what’s true, The Quantified Self makes the case that science belongs to everyone—and proves it by putting the method in the reader’s hands.

Wolf teaches ten reasoning patterns adapted from the history of scientific practice and illustrates them with dozens of personal narratives. Across chapters on pain, mood, cognition, sleep, digestion, heart health, and hormones, readers learn a practical method for answering the health questions that matter most to them…(More)”.

Article by Stefaan Verhulst and Andrew Zahuranec: “What would it take for Indigenous communities not simply to be represented in the emerging AI ecosystem but to shape, govern, and benefit from it?

That question is what animated our event during UNGA High-Level Week on Advancing Public Interest AI Through Data Commons for Indigenous Languages and Cultures. 

The discussion brought together Indigenous leaders and technologists, funders, researchers, international organizations, civil society, and industry around a growing challenge: As AI becomes embedded in education, public services, cultural production, and everyday life, many Indigenous communities are being drawn into AI systems that they did not design, using data they may not control, and reflecting assumptions and values that may not be their own.

The stakes are particularly high for Indigenous languages and cultural knowledge. AI could become a powerful instrument for language revitalization, education, cultural transmission, and economic opportunity. But without different approaches to data, infrastructure, governance, and investment, it could just as easily reproduce longstanding patterns of extraction while amplifying misinformation. This leads to weaker foundations of reliable information for AI and other systems that subsequently harm communities…Five themes emerged from the discussion…(More)”.