Stefaan Verhulst

Article by Sarah O’Connor: “In the early 20th century, a gifted engineer called Frederick Winslow Taylor embarked on an ambitious task: to extract knowledge from inside the heads of workers on America’s factory floors.

In the eyes of Taylor, who would go on to become one of the world’s first management consultants, factory workers possessed a “mass of rule-of-thumb or traditional knowledge” which had been “handed down from man to man by word of mouth” or “almost unconsciously learnt through personal observation”. Taylor thought it was about time this knowledge was “codified or systematically analysed or described”. To that end, he sent managers with stopwatches and notebooks on to shop floors to observe, time and record every stage of every job.

More than a century later, employers of white-collar professionals are beginning to confront a similar challenge. It is becoming increasingly clear that the knowledge required to make AI models genuinely powerful in a swath of workplaces is currently locked inside employees’ heads.

This isn’t true in every workplace. AI models have transformed the software profession, for example, because the task of writing code is testable and rules-based and there were vast reams of training data publicly available, thanks to online forums like Stack Overflow.

But for many other jobs, that sort of data just does not exist on the web. Indeed, some subtle but important skills are very hard to codify at all, which is why they are often learnt through experience and osmosis. This sort of tacit knowledge was famously summed up by the scientist and philosopher Michael Polanyi with the phrase: “we can know more than we can tell”.

As a result, general-purpose AI models are simply not very good at many specific tasks which require both domain and institutional knowledge. Investment firm Bridgewater Associates recently experimented, for example, with using LLMs to do something their human professionals do all the time: parsing reams of news stories and financial documents for information that might be relevant to their investment decisions.

While this could be a useful timesaver, Bridgewater found that variants of Gemini, Claude and GPT only tended to get it right about 50 per cent of the time..(More)”.

Article by The Economist: “WHEN THE leaders of the biggest Western economies gathered last month in the French Alps for the G7 summit, it was not just presidents and prime ministers in attendance. The bosses of America’s leading artificial-intelligence firms—Sam Altman of OpenAI, Sir Demis Hassabis of Google DeepMind and Dario Amodei of Anthropic—were seated alongside the political bigwigs. Even the stewards of the mightiest economies, it turns out, are buttering up the potentates of AI.

Days before the summit, America’s government had barred Anthropic from making its most advanced model available to foreigners. OpenAI soon imposed similar restrictions on some of its frontier systems. The move turned a simmering concern into a burning one: might countries without their own, home-grown AI firms end up excluded from the world-changing technology? Emmanuel Macron, France’s president, warned that no one would buy American AI if it could “turn off the switch” at will. The American AI labs’ Chinese rivals offer only faint reassurance. On July 7th Reuters reported that Chinese officials, too, were considering restricting foreigners’ access to leading AI models.

Sovereign chic

Caught between the world’s two AI superpowers, governments are increasingly pinning their hopes on “sovereign AI”: the idea that countries should limit their dependence on foreign providers of AI by actively developing domestic AI firms and infrastructure. Announcements of such initiatives are proliferating. The European Union has unveiled a technology-sovereignty package encompassing semiconductors, AI and cloud computing. Canada has launched its own plan to reduce reliance on American technology providers. India, Japan and Singapore are backing domestic AI infrastructure and models. The Centre for a New American Security (CNAS), a think-tank in Washington, estimates that the number of state-backed AI projects outside America and China grew fivefold over the course of 2024 and 2025 (see chart 1). Governments have announced investments of $70bn-plus in such initiatives, according to CNAS’s tally.

Yet turning these ambitions into reality will be extremely difficult. No matter what, most countries will still rely on American chips, Chinese open-source models or both to build their AI systems. Even then, most governments will find it prohibitively expensive to achieve anything remotely resembling AI sovereignty.

The incentive to seek it, however, is clear and compelling. Pablo Chavez of CNAS puts the restrictions on Anthropic in a “long line of events” demonstrating America’s willingness to use its technological dominion as a weapon. In the dying days of Joe Biden’s presidency in 2025, America proposed sweeping export controls that would have rationed access to advanced chips through a complex licensing regime. When Donald Trump became president a few weeks later he dropped the plan. But he has not hesitated to use America’s lead in AI as leverage in trade negotiations. The Department of Commerce is considering rules that would allow it to vet all sales of AI chips designed by American firms anywhere in the world. It might also insist that buyers allow inspections or monitoring to make sure chips are being used as promised. Another possible requirement could be to invest in American AI infrastructure in exchange for access…(More)”.

Policy Backgrounder by The Conference Board: “On the eve of the 2024 presidential election, while polls showed a race that was too close to call, online platforms called “prediction markets” told a different story. Contracts trading on the largest prediction market, Polymarket, placed the President’s odds of victory as high as 67% in the closing week of the race. For supporters of prediction markets, the President’s eventual victory – with nearly 58% of electoral votes – was evidence of their power to more accurately predict future events than existing methods, a phenomenon that has led Polymarket’s CEO to call them “global truth machines.” The outcome also helped shift what had once been a relatively obscure interest of some economists and traders into a mass market phenomenon – monthly volume on Polymarket and Kalshi jumped from less than $1 billion in mid-2024 to nearly $24 billion by early-2026.

Fundamentally, prediction markets are simple – users trade contracts tied to the outcome of a future event. In the most common format, an “event contract” might pay $1 if a candidate wins an election, if a hurricane makes landfall in a specified region, if inflation exceeds a stated level, or if a sports team wins a championship. This functionality could have both entertainment value – in the case of a contract that hinges on the length of the Super Bowl halftime show, for example – and real economic value – such as contracts predicting US gas prices.

For events with significant trading volume, supporters argue that these markets can translate large amounts of complicated real-world information into a market price that more accurately predicts events than existing methods such as polls or expert analysis. However, prediction markets have also raised significant regulatory and legal concerns related to alleged insider trading, outcome manipulation, regulatory arbitrage, and other issues. These questions have prompted policymakers and stakeholders to debate whether legal or regulatory action may be needed to protect users, clarify the boundary between trading event contracts and gambling, and preserve the potential economic value of prediction markets…(More)”.

Paper by Sachit Mahajan: “AI-assisted consultation can speed large-scale public engagement, but concise summaries may reflect some submissions more closely than others. This paper introduces participatory provenance, a framework for auditing how semantic coverage is distributed from submissions to summary sentences. Applied to two topics in Canada’s 2025 AI Strategy consultation (5,253 records; 2,861 participants), official summaries had higher observed mean coverage than exact-length random text, although statistical significance depended on the embedding model. Low coverage concentrated in semantic regions, especially those centered on criticism of educational technology and distrust of technology and oversight, whereas few or no records crossed the operational threshold in several better-covered regions. Same-budget, cross-fitted extractive benchmarks improved mean and lower-tail coverage on held-out submissions, showing that better semantic coverage was feasible without longer summaries. Consultation summaries should be evaluated not only for coherence and factual support, but also for how coverage is distributed across the range of submitted views…(More)”.

(Open Access) Book by Regina Lenart: “…explores how public organizations can effectively design, implement and manage crowdsourcing initiatives… highlights the vital role played by crowdsourcing in fostering openness, inclusivity and transparency in the public sector, despite rapid digital developments.

Bridging theory and practice, this interdisciplinary book presents the current trends, tools and techniques for crowdsourcing, using international case studies to demonstrate crowdsourcing applications across a variety of economic, political and cultural contexts. The book looks to the future of crowdsourcing, examining the role of digital tools and citizen engagement in public decision-making, policy design and social problem-solving…(More)”.

Article by Terrence O’Brien: “The US Department of Energy reportedly deleted about 6,000 pages related to energy conservation as a historic heatwave tears across the country.

The deletion was suspiciously timed, following Republican outrage over Mayor Zohran Mamdani asking New Yorkers to help reduce strain on the grid by setting their AC to 78 degrees. Republicans like Ted Cruz (who has famously fled severe weather in his home state), Nikki Haley, and Representative Nancy Mace (South Carolina) quickly pounced, framing the request as socialism and an act of war on women in menopause (the Republican Party is notoriously concerned about women’s health).

Of course, this is pretty standard advice during a heatwave. It was the official stance of the Department of Energy that Americans should set their thermostats between 75 and 78 degrees, and Republican governors in deep red states like Texas have issued the same advice in the past — including current governor Greg Abbott.

The deletions by the Trump administration are broad and indiscriminate. While pages that would support Mamdani’s request to lower thermostats were deleted, so too were pages about water conservation, types of insulation, and its solar decathlon challenge. The Internet Archive preserved the pages that have been lost…(More)”.

Article by Stefaan Verhulst, Nadiya Safonova and Hannah Chafetz: “The world is facing increasingly complex challenges, including higher levels of conflict, displacement, political polarization and social fragmentation. Addressing these challenges requires new tools and approaches that can support conflict prevention, peacebuilding, and advancing resilient societies. 

PeaceTech is the intentional use of technologies and data to save lives, safeguard human dignity, prevent, mitigate, or recover from conflict, enable accountability, and help people live with dignity, agency, and security. 

Yet while enormous investment has gone into technologies of war, comparatively little attention has been devoted to technologies explicitly designed to prevent violence, protect civilians, strengthen resilience, and build lasting peace. This mapping seeks to help fill that gap.

Drawing on desk research conducted between February and May 2026 and learnings from supporting the 2023 – 2025 Kluz Prize for PeaceTech, we identified and curated 100 current and potential use cases of PeaceTech across the conflict cycle and for long-term peacebuilding. 

This mapping is intended to be illustrative rather than exhaustive. Given the rapid pace of technological innovation and the diversity of peacebuilding contexts, no single review can capture every current or emerging PeaceTech application. Instead, the 100 use cases are designed to demonstrate the breadth of possibilities, identify promising patterns, and stimulate further discussion, research, and innovation across the field.

More detailed case studies about these use cases along with the potential and risks of PeaceTech will be published in our forthcoming white paper this fall. 

The 100 Use Cases

Figure 1. Screen capture of 100 Use Cases of PeaceTech

The 100 Use Cases are categorized by maturity level. As shown in the above spreadsheet, we identified: 

• 20 established use cases that have been deployed at scale with evidence of impact; 
• 20 growing use cases that have active pilots or are in the early stages of deployment;
• 20 use cases in the conceptual stage of design; and 
• 40 unexplored use cases that have potential but have not yet been fully established or deployed. 

In what follows we outline several key themes across each of these categories. Three broader observations emerged from reviewing more than one hundred current and future examples:

• Most PeaceTech today focuses on responding to conflict rather than preventing it.
• AI starts to appear across nearly every stage of the conflict cycle.
• The largest innovation gap lies not in hardware but in coordination, governance, incentives, and trusted institutions…(More)”.

Paper by Stefaan Verhulst: “Across a range of fast growing urban markets, private developers are constructing a version of the smart city that operates largely outside the purview of municipal government, often at the explicit invitation of city officials seeking to shift the cost and complexity of digital infrastructure onto private capital. Gated residential and mixed use developments are increasingly marketed not merely on the basis of security and amenity, but on their smartness: integrated home automation, app mediated access control, and centralized energy and resource management, among other features. We refer to this phenomenon as the smart compound. Despite its rapid proliferation, it has received comparatively little sustained scholarly attention: the literatures on smart cities and on gated communities have developed largely independently of one another, even as developers are, in practice, merging the two. This paper introduces the smart compound as an emerging real estate and urban development phenomenon, considers the opportunities and risks it presents, and examines how questions of data governance differ when smart urban infrastructure is built and owned privately rather than publicly. It concludes with a set of research questions intended to orient researchers, planners, and regulators toward a phenomenon whose growth is outpacing the scholarship meant to account for it…(More)”.

Paper by Francesco Nasi: “While many debates frequently emphasize the threats artificial intelligence poses to democratic life, scholars are increasingly examining the role AI may play in supporting democracies. However, little attention has been given to a crucial question: how and under what conditions AI can be considered democratic, not only in its design but also in its societal outcomes. I argue that AI is democratic when it is democratically empowering, meaning it contributes to redistributing power and creating symmetrical power relations rather than centralizing them. This argument is rooted in the framework of participatory democracy, which prioritizes the active involvement of the citizenry and the redistribution of power across all social spheres. Drawing on participatory democracy, Foucauldian and Actor-Network Theory perspectives, the paper identifies three essential features of power that are needed for theorizing democratic AI: the pervasiveness of power relations in everyday life, the agency of technological artifacts understood as “actants”, and the way power structures relations that can be more or less symmetrical. To assess whether and how AI may be deemed democratic, the paper suggests investigating how AI shapes power dynamics in two dimensions: AI making (development, design, political economy, governance, and imaginaries) and AI action (how AI operates in fields like education, healthcare, information, and politics). This approach helps to take into account the complexity and ambiguity of power dynamics, critically examining whether and how AI contribute to the redistribution of power across different domains. Ultimately, this theoretical framework provides a roadmap for developers and policymakers interested in promoting a more democratic AI…(More)”.

Article by Daniella Fernández, and Andrea Paola Hernández: “After twin earthquakes struck northern Venezuela last month, thousands of people took to social media, pleading for help locating friends and family. The interim government was slow to act, but help came quickly from an unexpected source: developers and programmers.

“Who knew when the government was going to respond?” Jorge Bastidas, a 31-year-old Venezuelan programmer who lives in Buenos Aires, told Rest of World. “We decided to take action.”

Bastidas and his team of six created Desaparecidos Terremoto Venezuela, a website for citizens to report, identify, and reunite with missing family and friends. Using Claude Opus 4.8, and facial recognition software donated by Mexican company Lab-Co, Bastidas designed the site to load quickly, without requiring users to register or download an app. It received more than 30,000 missing-person reports in the first two days, he said.

Without AI, “it would have taken me about 24 hours without rest to build something that took me three hours,” Bastidas said…(More)”.