Stefaan Verhulst
Article by David Adam: “When psychologist Raluca Rilla asked volunteers to complete a survey last year, she got the following response to one of her questions: “I don’t experience confusion in the same way humans do.”
Rilla, a PhD student at the Max Planck Institute for Human Development in Berlin, suspects that this is the obvious tip of a large and worrying iceberg — one that could scupper academic research on how people think and behave. She and her colleagues estimate that up to 45% of responses they receive to such surveys are now copied and pasted from the output of large language models (LLMs). In some cases, participants might simply be polishing their language. In others, Rilla thinks that the entire operation — signing up, reading the questions and submitting responses — is handled by a machine. Such answers, and the academic studies built on them, are unlikely to reflect the reality of human nature.
Experimental psychology is not alone in wrestling with the impact of LLMs on research. From political science and economics to opinion polling, researchers across the social sciences are sounding the alarm after finding the fingerprints of artificial intelligence and considering the implications. AI chatbots are infiltrating social-science surveys — and getting better at avoiding detection
Even if AI input into polls can be throttled, there’s a concern at the analysis stage, says David Lazer, a political and computer scientist at Northeastern University in Boston, Massachusetts: AI-assisted analyses in social science might flood journals with spurious findings by rapidly whipping up studies. One journal has already chronicled a vast increase in the number of manuscripts it has received that were wholly or mostly prepared using AI tools.
The explosion in the use and power of AI models touches researchers across all academic fields. But the impact on the social sciences is especially acute, says Joshua Tucker, a political scientist at New York University. That’s because, compared with other disciplines, much social-science research is heavily reliant on survey data and analysis. And when researchers aren’t gathering the data themselves, they are often analysing large, general data sets, such as censuses or other huge surveys that were usually collected for a different original purpose. This means that apparent signals in the data can be plucked from noise in a way that isn’t possible with experimental data obtained in narrow tests to check a hypothesis — information that tends to have a single use and a defined shelf life.
“I think we’re approaching a time where the trust in behavioural and social sciences will be undermined by this constant threat of LLM pollution,” says Björn Hommel, a psychologist at Leipzig University, Germany. “And there’s nothing that we are able to do about it right now.”
But it’s not all doom and gloom. An alternative view of the latest AI systems is that they could transform social science by making its findings more robust. The same algorithms that can be used for superficial work such as polishing language can also source and analyse complex data sets quickly and, by toggling through statistical techniques, check how sensitive an individual finding is to various analytical methods. AI-assisted review could help to spot methodological errors, and social-science journals might insist on the use of more-robust methods as AI makes it easier for researchers to attempt them…(More)”.
Article by Eric Niiler: “The Trump administration is dismantling a $368 million deep-ocean observation system that was put in place a decade ago to monitor coastal environments, marine ecosystems and powerful currents that affect the global climate.
The National Science Foundation said it would send ships in June to begin removing more than 900 deep-sea instruments anchored off Oregon, Washington State, Alaska, North Carolina, and an area between Greenland and Iceland known as the Irminger Sea.
Scientists have used data from the system to understand how the ocean is absorbing greenhouse gases from the atmosphere, how changes in ocean temperature such as marine heat waves might affect fisheries or signal bigger shifts in the climate, and coastal flooding along the East Coast…(More)”.
Roadmap by Centre for Social Impact: “How do you measure the real difference your work is making for people and communities – particularly when social issues are interconnected, persistent and solutions are constantly evolving?
For many organisations, social impact measurement can feel challenging. While it’s often easy to track activities or outputs, understanding whether meaningful change is happening over time is far more difficult.
The Roadmap to Social Impact is a practical guide to social impact measurement, designed to help you better plan, measure and communicate the change you’re helping create.
We hope it provides both guidance and inspiration as we work together for a better world…(More)”.
Article by Juan Ortiz-Freuler: “This article reveals significant market consolidation within the digital infrastructure of Latin American media. Extending the political economy of media literature from ownership to control over underlying infrastructure, it uses a newly constructed database of over 400 data points to analyze the online infrastructure of 18 media outlets across 6 Latin American countries, focusing on 11 key elements of the media stack. Findings indicate that dominant players, such as Alphabet and Meta, pose the greatest commercial risks to the analyzed media. Meanwhile, the U.S. government emerges as the greatest geopolitical risk, with 50–100% of providers across the analyzed elements operating under U.S. law, exposing Latin American media with similar infrastructural profiles to U.S. government policy. The article places these challenges in conversation with historical calls for a New World Information and Communication Order, which underlines that a robust understanding of media autonomy requires infrastructural autonomy…(More)”.
Paper by Global Solutions Initiative: “…introduces “plural protocol ecosystems” as a comprehensive alternative to the centralized, extractive digital infrastructures currently dominated by massive tech monopolies. It argues that while existing European regulatory approaches like competition law try to govern the behavior of digital gatekeepers, they fail to alter the architectural logic that causes power and value to concentrate in the hands of a few. By shifting focus away from single corporate-led platforms and toward community-led, open digital protocols—the fundamental rules determining how digital systems interact—the framework seeks to build a value-driven third path for the digital age. This new infrastructure aims to actively embed democratic values, human rights, and strategic sovereignty directly into technical designs rather than trying to enforce them after the fact.
To prevent digital systems from collapsing into winner-takes-all dynamics, the paper outlines a framework built across three deeply interconnected dimensions: technical, economic, and social. Technical plurality ensures structural resilience and neutrality through open-source code, decentralized architectures, and privacy-by-design. Economic plurality re-engineers value flows to create fair, non-extractive distribution systems where data creators, developers, and users are directly compensated for the value they generate. Finally, social plurality establishes participatory, multi-stakeholder governance models that prioritize broad deliberation, consensus across social differences, and the right for communities to exit or fork a protocol if it no longer serves them.
Ultimately, the paper positions this paradigm as a path toward a modern digital enlightenment, transitioning individuals from passive consumers of centralized systems into active co-creators with genuine agency over their digital lives. By satisfying these foundational technical, economic, and social blocks, a digital ecosystem can naturally foster higher-level qualities like verifiable trust and credible neutrality. This ground-up architecture provides a realistic roadmap for democratic powers to cultivate a competitive, resilient, and sovereign digital foundation that honors public interest and the common good…(More)”.
Paper by Robert Axelrod and Scott E. Page: “Sycophants praise and support leaders’ proposals to gain personal and professional advantage. A rational sycophant is an advisor who supports actions they expect to be harmful even when rewards and punishments for good and bad advice are equal in magnitude. Rational sycophancy arises when the outcome distribution of a proposed action has a negative expected value but a positive median (outcome asymmetry). The risk is greatest when a small yet meaningful fraction of outcomes are catastrophic, which occurs in long-tailed distributions. Given that single realizations from long-tailed distributions reveal little about the underlying distribution, even after outcomes are observed, a leader may be unable to distinguish rational sycophancy from wise counsel. As a result, rational sycophants may gain influence and increase the likelihood of catastrophic policy outcomes…(More)”
Article by Alice Xiang: “Every click, every photo, every search query we make creates a digital echo. These digital traces are the raw material fueling the AI revolution, powering technologies that are reshaping our world. Yet for the people creating it—all of us—this data has become functionally worthless.
The average internet user doesn’t think about the value of their data. They simply give it away to some of the wealthiest companies in the world, for free.
Because of this behavior, I fear we are living in an age of data nihilism, where our data means everything to AI developers yet almost nothing to us—not because our data actually is value-less, but because people feel powerless to stop it from being collected against their will. …
The next step is for the AI community and regulators to take ethical data curation seriously. The economic power dynamics between AI and humans will largely be determined at the data layer, and as a result, questions about consent and compensation mechanisms for data rights holders should be a major area of focus for AI researchers and regulators. Creating opt-in or opt-out schemes that provide meaningful control to people around the world whose data serves as AI’s raw materials is a challenging task, but one that is critical to address now. Moreover, as AI developers exhaust available data, future innovations will likely depend on the quality rather than simply the quantity of data.
Nietzsche’s cure for nihilism was to create personal meaning, but the scale of AI necessitates creating systems that affirm and protect the value of humanity’s contributions. We are now at a turning point: if we fail to build such protections, we will resign ourselves to a future where the benefits of AI are concentrated among a few, and the vast majority of people find their contributions worthless. The future of AI should not be built on a foundation of mass data appropriation. It must be built on a foundation of respect, consent, and shared value. The age of data nihilism is upon us; it is up to us to prevent it…(More)”.
Report by the Tony Blair Institute: “Modern states increasingly depend on digital infrastructure that is both critical and physically vulnerable. Identity systems, registries, payments, legal records, administrative platforms and public-service channels now form part of the state’s operational core. When these systems fail, the issue is not only service disruption, but also the continuity and credibility of government itself.
War, cyber-attacks, sabotage, terrorism and natural disasters can all affect the systems on which governments, economies and public services rely. Artificial intelligence intensifies this challenge. Governments will increasingly depend on concentrated, capital-intensive infrastructure – cloud, compute, energy, connectivity, secure data environments – that many states cannot fully build, finance or control domestically. As a result, the question is no longer whether states will depend on infrastructure beyond their borders, but how that dependence can be governed.
One concept attracting growing attention as a potential solution to these challenges is the digital embassy – a legally, technically and politically governed arrangement that allows states to preserve, restore or operate critical digital functions through trusted infrastructure beyond their territory. Building on the earlier data-embassy model, which focused on legally protected cross-border storage and recovery of critical state data and systems, digital embassies go even further. They can provide live-system continuity, service continuity and, in some cases, trusted access to cloud, compute or AI infrastructure. At their most advanced, digital embassies are designed to preserve the operational capacity of the state – including identification, authentication, access to records, legal administrative processes and communication with citizens – even when domestic infrastructure or normal administration is disrupted.
This paper argues that digital embassies should be understood as strategic instruments for resilience, security, sovereignty and capability. For guest states (those seeking to externalise elements of their sovereign digital infrastructure), they can reduce single points of failure, preserve state continuity and provide access to strategic infrastructure. For host states (those aiming to host such systems for other countries), they can attract investment, strengthen the economy, aggregate demand for cloud and compute, deepen strategic partnerships, and position countries within the emerging global digital and AI infrastructure ecosystem. But hosting also creates political, legal and security exposure, and therefore requires credible safeguards and institutional capacity…(More)”.
Report by Oliver Escobar and Stephen Elstub: “Democratic wellbeing means people having voice and choice over decisions that affect their lives. Strengthening parliamentary democracy is essential for improving democratic wellbeing in the UK. This paper supports current efforts to enhance the UK’s four parliaments through deeper public engagement and citizen deliberation. Building a stronger working relationship between parliamentarians and citizens can increase institutional capacity to address the challenges of our time.
Parliaments are expected to navigate complex, long-term and often contentious policy challenges, while maintaining public trust and demonstrating responsiveness to citizens. Yet existing mechanisms for engaging the public in parliamentary policymaking – consultations, surveys, lobbying – often struggle to provide the kind of considered public judgement these challenges require
Why mini-publics?
We focus on mini-publics because they provide a robust and practical way to embed citizen deliberation in legislatures. Mini-publics are civic forums in which a diverse cross-section of the public participates in evidence-based deliberation to inform decision-making. Participants are selected through civic lottery to reflect the wider population and supported with resources to reduce barriers to participation…(More)”.
Book by Keith McCandless and Nancy White: “Liberating Structures (LS) are simple patterns that make it possible for individuals and groups to work together brilliantly. With a menu of 43+ open-source patterns, LS shift how people meet, plan, decide, and relate to one another, making it possible to include and engage many voices in shaping next steps and the future. LS Their minimal set of structural constraints to set groups free from conventional patterns that exclude, stifle, and over-control (like endless presentations, managed discussions, or unfocused brainstorming)…What’s New: Inside the Fieldbook
Sparkling New Visuals: Every structure features a sharp visual format and simple line drawings, making the microstructures and templates easy to follow, adapt, and share.
Expanded Repertoire: The LS menu has grown to 43 methods, introducing ten new, second-generation Liberating Structures to tackle modern challenges.
Integrated Online Practice: No more fragmented guides. Practical, field-tested instructions for both online and face-to-face facilitation are now embedded directly within every single structure.
Inclusive Design & Continuity: Learn how to co-design with your group. Master the art of composing strings, using short punctuations, and tracking multi-session progress with your participants, lifting off precisely where you left off.
The Deeper Why: We shine a brighter light on the 10 LS Principles and the complexity science concepts that underpin the repertoire—helping you bridge the gap between your espoused values and routine habits…(More)”