Stefaan Verhulst

Report by Katherine Barrett and Richard Greene: “State governments are increasingly exploring how GenAI can streamline operations, enhance service delivery, and support policy innovation—while safeguarding human judgment, transparency, and accountability that define public governance.

Through an in-depth review of current pilot projects, emerging use cases, and early implementation lessons, the authors offer a forward-looking perspective on how GenAI can serve as a collaborative partner for state employees. The report maps areas where AI can complement, augment, or automate tasks within diverse state functions, from public health and transportation to education and environmental management.

Key recommendations include fostering cross-agency learning networks, investing in targeted workforce training and upskilling, and adopting governance frameworks that balance innovation with responsible use. By following these strategies, states can cultivate a technologically empowered and resilient workforce in an era of rapid digital change…(More)”.

Article by Jenny Gross: “…About 1,500 letters are sent once a year to randomly selected residents in Ostbelgien. Of those who indicate interest, about 30 are chosen to join the citizens’ assembly.

Starting in September, they meet on Saturdays for several hours over a period of two months, or longer if needed, and are assigned a topic. Each participant is paid a stipend of about 115 euros ($133) per day. They gather in the regional parliament building, which served as a military hospital during World War II, with a moderator employed by the government facilitating the discussions.

Though the assemblies’ recommendations are not binding, lawmakers are required to consider them, and many have been adopted. Among the changes they have spearheaded: easing eligibility requirements for low-income housing; including residents’ family members on the boards of assisted-living facilities; and new funding to encourage young people to take up professions such as nursing, which is facing a shortage in the region.The Belgian experiment recalls ancient Athenian democracy, in the 5th century B.C., when groups of free men were chosen at random to serve as government officials each year. There wasn’t much diversity in that citizenry, however, and these days, leaders in Eupen, the capital of Ostbelgien, acknowledge that what works in their small, relatively homogenous region may not translate everywhere.

The assemblies’ purview is also limited, naturally, to areas where the regional government has control, such as education and housing, rather than more divisive topics like the entry of immigrants which is overseen by the federal government in Brussels…(More)”.

Paper by Megan A Brown et al: “Scientists across disciplines often use data from the internet to conduct research, generating valuable insights about human behavior. However, as generative artificial intelligence relying on massive text corpora becomes increasingly valuable, platforms have greatly restricted access to data through official channels. As a result, researchers will likely engage in more web scraping to collect data, introducing new challenges and concerns for researchers. This paper proposes a comprehensive framework for web scraping in social science research for U.S.-based researchers, examining the legal, ethical, institutional, and scientific factors that we recommend researchers consider when scraping the web. We present an overview of the current regulatory environment impacting when and how researchers can access, collect, store, and share data via scraping. We then provide researchers with recommendations to conduct scraping in a scientifically legitimate and ethical manner. We aim to equip researchers with the relevant information to mitigate risks and maximize the impact of their research amid this evolving data access landscape…(More)”.

Paper by Erick Elejalde et al: “This study examines behavioral responses after mobile phone evacuation alerts during the February 2024 wildfires in Valparaíso, Chile. Using anonymized mobile network data from 580,000 devices, we analyze population movement following emergency SMS notifications. Results reveal three key patterns: (1) initial alerts trigger immediate evacuation responses with connectivity dropping by 80% within 1.5 h, while subsequent messages show diminishing effects; (2) substantial evacuation also occurs in non-warned areas, indicating potential transportation congestion; (3) socioeconomic disparities exist in evacuation timing, with high-income areas evacuating faster and showing less differentiation between warned and non-warned locations. Statistical modeling demonstrates socioeconomic variations in both evacuation decision rates and recovery patterns. These findings inform emergency communication strategies for climate-driven disasters, highlighting the need for targeted alerts, socioeconomically calibrated messaging, and staged evacuation procedures to enhance public safety during crises…(More)”.

Article by Roeland Beerten, Johannes Jütting and Stefaan G. Verhulst in Le Monde: “Les statistiques officielles – fondement d’une gouvernance fondée sur des faits – sont aujourd’hui prises entre deux feux : la politique et la défiance du public. Dans certains pays, les agences sont marginalisées ; dans d’autres, les citoyens doutent de leur utilité. Si les systèmes statistiques ne parviennent pas à refléter les réalités vécues, s’ils ne fournissent que de simples moyennes abstraites, ils risquent de devenir une victime de plus de la crise de confiance démocratique.

« L’inflation est maîtrisée. » « Le PIB [produit intérieur brut] progresse. » « Nos villes sont sûres. » Ces affirmations, répétées à longueur de communiqués, suscitent désormais plus de méfiance que d’adhésion. Car pour de nombreux citoyens, les factures s’envolent, les économies s’amenuisent et l’avenir des générations futures s’obscurcit. Quand les chiffres contredisent l’expérience vécue, la confiance dans les experts s’effondre.

Depuis des décennies, les offices statistiques décrivent le monde à coups d’agrégats : PIB, chômage, inflation. Ces indicateurs sont utiles, mais leur logique gomme les écarts : le macro prime sur le micro, la moyenne efface la marge. Dire que « l’économie croît » n’a guère de sens pour celui qui voit son pouvoir d’achat s’éroder. D’où la montée d’une exaspération, que l’on pourrait résumer ainsi : « On ne mange pas du PIB. »..(More)”.

Blog by Anna Colom, Marta Poblet, and Stefaan Verhulst: “Independent and public interest media have for long been considered a key public good for societies, a pillar of democracy and accountable governance. Well funded public media systems have been found to consistently correlate with ‘healthy democracies’. However, the struggle for public interest media as a resilient sustainable public good is a long-standing one. Decades-long market fragmentation, the digital transformations turning the public sphere into multimodal networked spheres, and more recently, the entering into the mainstream of generative AI (GenAI) systems, and General Purpose AI more broadly, have kept public interest and independent media in a constant state of survival and reinvention.

The current trends in AI development -particularly large GenAI models- risk creating a vicious cycle for journalism and public interest media. These models are trained on high-quality journalistic content but the methods for capturing the data have been associated with exploitative and extractive practices of web scraping, routinely disregarding property rights, licensing, quality, (mis)representation and bias as GenAI models often use these data without providing provenance, context or integrity. Once deployed, they often draw audiences away from the very outlets that produced this content by providing AI-generated summaries and news-like outputs directly to users. As a result, media organisations lose readership, along with the advertising and subscription revenue that sustains independent reporting. With fewer resources, their ability to produce quality journalism declines, which in turn reduces the availability of trustworthy content. This cycle threatens not only the financial viability of public-interest media but also the integrity of the AI systems that depend on their work, and ultimately, the integrity of the available knowledge to the public…(More)”.

Article by Anjana Ahuja: “…The Genuinely Hard Problems scheme, designed to expose bright young minds each week to the world’s biggest unanswered questions, might usefully chart a course for other institutions to follow. According to Logan McCarty, a Harvard science lecturer and dean of education who is organising the classes with the scheme’s creator, neurobiology professor Jeff Lichtman, the internet and AI have lessened the need for ambitious thinkers to acquire specialised technical skills and internalise vast quantities of information…

Specialist knowledge can now be digitally retrieved in seconds; AI can mine data, construct hypotheses and design experiments. On top of that, a slender scholarly lens can obscure a wider perspective. Today, some of the biggest problems facing humanity, such as climate change and energy scarcity, tend to sprawl across disciplines rather than sit snugly within academic departments.

The primary task of scientists, the Harvard educators believe, is asking the right questions, because AI can answer even difficult queries if they are well-posed; being fearless and willing to fail, with no area of science off-limits; and doing research that is meaningful and has impact, rather than chasing quick wins…(More)”.

Book by Susan C. Stokes: “Democracies around the world are getting swept up in a wave of democratic erosion. Since the beginning of the twenty-first century, two dozen presidents and prime ministers have attacked their countries’ democratic institutions, violating political norms, aggrandizing their own powers, and often trying to overstay their terms in office.

The Backsliders offers the first general explanation for this wave. Drawing on a wealth of original research, Susan Stokes shows that increasing income inequality, a legacy of late twentieth-century globalization, left some countries especially at risk of backsliding toward autocracy. Left-behind voters were drawn to right-wing ethnonationalist leaders in countries like the United States, India, and Brazil, and to left-wing populist ones in countries like Venezuela, Mexico, and South Africa.

Unlike military leaders who abruptly kill democracies in coups, elected leaders who erode them gradually must maintain some level of public support. They do so by encouraging polarization among citizens and also by trash-talking their democracies: claiming that the institutions they attack are corrupt and incompetent. They tell voters that these institutions should be torn down and replaced by ones under the executive’s control. The Backsliders describes how journalists, judges, NGOs, and opposition leaders can put the brakes on democratic erosion, and how voters can do so through political engagement and the power of the ballot box…(More)”.

Report by Careful Industries: “…discusses what we learnt about how local communities and everyday publics make sense of data-intensive and AI-based technologies in their everyday environments.

Through engagements with participants in four locations across the UK and at one site in Australia, our research found that there is a persistent perception that government discourse on the societal benefits of AI is disconnected from the needs of local communities in the urban environments where AI innovation takes place.

The report reflects on future opportunities for public participation in AI governance and offers recommendations aiming to grow public trust and deliver better outcomes…(More)”.

European Union: “The strategy identifies three priority areas for action based on:

1. Scaling up access to data for AI to ensure our businesses have access to high-quality data needed for innovation
2. Streamlining data rules to give legal certainty to businesses and reduce compliance costs
3. Safeguarding the EU’s data sovereignty to strengthen our global position on international data flows

Scaling up access to data for AI

Flagship initiatives to address data bottlenecks:

• Launch the first data labs to scale up data availability and create links between data spaces and AI ecosystems – they will pool both private and public resources to make high-quality sectoral data available to companies, including SMEs, and researchers using AI and provide relevant data services for AI-driven innovation
• Scale up common European data spaces, supported by ongoing EU investment of around EUR 100 million, creating new data spaces across key sectors, including a defence data space.
• Explore horizontal enablers to boost the entire data economy. In particular, expanding high-value datasets under the Open Data Directive, making 30 million digitised cultural objects available for AI training, boosting the use of synthetic data and the EU’s production output…(More)”.