Louis Ravn’s Post

🧪 Synthetic Data Meets Development: Justice or “Synthetic-Washing”? New in Big Data & Society: “Towards synthetic data justice for development: A case study of synthetic datasets on human trafficking” by Louis Ravn Synthetic data is increasingly promoted as a privacy-safe solution for working with sensitive issues—such as human trafficking. But does it truly eliminate harm, or does it risk masking deeper inequalities? This paper investigates the UN–Microsoft collaboration that produced the Global Synthetic Datasets (GSDs) on human trafficking—now representing over 222,000 simulated survivor profiles. Through a justice lens, the article raises critical questions: Political Economy: Who controls synthetic datasets, and who benefits? (In)Visibility: Can synthetic data make marginalized groups visible without deep misrepresentation? Abstraction vs. Experience: What is lost when trauma becomes “synthetic”? “Synthetic-Washing”: The danger that privacy claims may obscure enduring power asymmetries. Ravn calls for a framework of synthetic data justice—one that goes beyond technical fixes to address ethical, political, and social realities in development contexts. 👉 Read the article: https://lnkd.in/ej74DeA4

Superb new publication by Louis Ravn on the increasingly popular role of synthetic #data in contemporary #development practice. Ravn calls for a framework of #synthetic data justice—one that goes beyond technical fixes to address ethical, political, and social realities in development contexts. It will be part of our forthcoming special issue on 'datafied development with Big Data & Society (Sage Journals), which also contains excellent papers from wonderful scholars on topics ranging from from #digital #humanitarianism to AI4D to the #neocolonially of #AI. Keep an eye out for collection. Nanna Thylstrup Julian P. Rafael Grohmann

The biggest blocker to ethical data use isn’t consent or encryption. It’s translation. When datasets can’t “speak”, good ideas die in handoffs. I’ve seen projects with funding, trust and purpose stall because teams spent months mapping columns, not solving problems. That’s why the UK Data Service shift to DotStat with SDMX matters. It gives us a shared language for structure and meaning. SDMX 3.1 landed this year, and it shows in the new Data Explorer experience. It isn’t a warehouse. It’s an interpreter that reduces ambiguity at source. What changes on the ground: → Local “country of birth” stats align cleanly with health research on migrants, without brittle one-off mappings. → New census tables can be compared with older series, because semantics travel with the data. → AI pipelines stop wrangling and start learning, thanks to consistent, machine-readable definitions. Under the hood, DotStat wires components so external sources can connect safely via a transfer layer. That means UKDS can view OECD structures in its own lifecycle manager without cloning everything. Interoperability without copy-paste sprawl. This is privacy by design in practice. Shared structure reduces re-collection, tightens governance, and makes FAIR the default, not an aspiration. The payoff isn’t just compliance. It’s faster questions answered, with less risk baked in. What’s next is encouraging: UKDS is moving recent census data into SDMX, then backfilling older series, so reuse becomes normal across programmes and time. That’s how you scale trustworthy research without scaling friction. If your team had a common data language tomorrow, which thorny join would you fix first and what impact would that unlock? #datainteroperability #SDMX #PrivacyByDesign #data #privacy

I’ve been working on something quite special over the past few months. Please find attached. I believe that synthetic data is the next fundamental step towards open innovation. It allows predictive modelling, data sharing, and collaborative research without compromising privacy. By enabling safe access to realistic data, we can empower teams - across organisations and disciplines - to work together more freely and accelerate discovery responsibly. The attached report presents my exploration of synthetic data generation using the PrivBayes method, as per my recent experiments and observations. It includes practical metrics for comparing real and synthetic datasets. The report is structured to be accessible to a broad audience: it begins with a high-level overview for readers new to the topic, and moves into more technical discussions for those interested in the mathematics, modelling, and evaluation methods. So whether you’re in data science, policy, or pure mathematics - there’s something here for you. My goal is to show that privacy and utility can coexist - that synthetic data can serve as the foundation for responsible, collaborative innovation without requiring security clearance or access to sensitive information. I’d love to hear your thoughts. Note also that PrivBayes is one of the models available in SAS Data Maker - a great tool that can help streamline synthetic data generation. #SyntheticData #OpenInnovation #DataScience #Privacy #Mathematics #MachineLearning #Collaboration #AI

New technology isn’t the problem — pretending data is neutral is. I’m honored to have my commentary featured in The Boston Globe: “New technology isn’t the problem. We need to stop pretending that data is neutral.” This piece is personal. I grew up inside the very systems I can now help shape — a kid labeled “troubled” before anyone asked why. Back then, decisions were made about me using data I didn’t understand and had no voice in. Now, as a youth advocate and data scientist, I see those same harms — biased datasets, “neutral” algorithms, and decisions made behind closed doors. The harm isn’t always loud, but it’s lasting. Technology isn’t the enemy. The danger lies in what we build, how fast we build it, and who we build it for. When ethics and equity are afterthoughts, injustice gets automated — at scale. If we don’t start with ethics, we end with harm. Real progress begins with listening to those closest to the problem and asking: Just because we can build it, should we? Here’s to building systems rooted in humanity — and data that tells the whole story. https://lnkd.in/eTDQ_4vP #DataEthics #YouthJustice #SystemsChange #TechForGood #Equity #Leadership #SocialImpact #BostonGlobe

Data – the oil of the 21st century? ⛽➡️💾 In 2017, The Economist declared that the world’s most valuable resource is no longer oil, but data. Fast forward to 2025, and we can now judge that statement from a fresh perspective. 2017: the early debate ➡️ Data became the fuel for AI algorithms. ➡️ More data → better products → more users → even more data. ➡️ Big Tech built moats almost impossible to cross. ➡️ Traditional antitrust rules looked powerless. 2025: today’s reality ➡️ EU fines: Apple, Meta, and soon Google – under the Digital Markets Act. ➡️ New regulations: the Data Act, European Health Data Space, and the UK’s Data (Use and Access) ➡️ Act 2025 are shifting data toward shared infrastructure. ➡️ Startups warn: data dominance is not just about technology, but also a strategy to block competition. ➡️ Privacy & transparency have become the core conditions for trust and competitive advantage. 👉 Today, data is no longer just a “resource”. It is strategic infrastructure – shaping the balance between innovation, privacy, and competition. 💡 Question for you: should data be treated as a public good (like roads or energy), or remain fully in the hands of Big Tech? ⚡ My bold take: if nothing changes, by 2030 the Big Five will control more of the global economy than all G7 nations combined.

Last week, in our open meeting with data scholars, when discussing potential research scope, a thoughtful comment from a colleague about #dataharm stayed with me all week. It prompted me to revisit a 2023 Information & Organization article, "Future directions for scholarship on data governance, digital innovation, and grand challenges", by Elizabeth Davidson, Lauri Wessel, Jenifer Sunrise Winter, and Susan Winter. The grand challenges outlined in that piece—and the special issue about #datagovernance —feel even more urgent today. Scholars studying digital innovation, data-intensive organizing, and AI in management more broadly, are so used to talking about the many benefits using data. But there are no guarantees that only desirable outcomes will develop; good uses exist only when the thoughtful governance is in place. It is true: contact-tracing apps supported early COVID-19 responses; and data from wearables, smart devices, and in-home monitoring can help families support aging at home. But these benefits always come with an “IF”—if systems are designed for adoption and trust, and if we address privacy, consent, and equity head-on. They call for meaningful approaches to data governance within and across organizations that not only facilitate digital innovation but also balance social, economic, and technical benefits and risks for individuals, organizations, and societies. What I value most is the multi-level agenda laid out in this paper—micro and macro level governance, new organizational forms, #opendata, equity and fairness in policy etc—offering a practical roadmap to steer digital innovation toward the public good. I strongly recommend it for anyone working on data/AI ethics, data governance, or responsible innovation. Here is a link to the paper: Davidson, E., Wessel, L., Winter, J. S., & Winter, S. (2023). Future directions for scholarship on data governance, digital innovation, and grand challenges. Information and Organization, 33(1), 100454.  https://lnkd.in/gSP4y6TJ We feature other papers focusing on data as an object of research in the Data Studies Bibliography at https://lnkd.in/gEfef6Rw . Take a look! #datastudies #data #research

Who will write the rules of the digital age: tech giants, states or citizens themselves? Jitendra, your post highlights how data supremacy has turned platforms into powers without borders. When information becomes currency, those who control it shape markets, influence choices and even frame beliefs. You are right that national sovereignty is being redefined under the weight of corporate power, with tech platforms increasingly influencing how societies authenticate, transact and communicate. What is striking is that this contest unfolds on multiple fronts. At the geopolitical level, governments are not standing still. Europe is pushing its own vision of digital fairness while countries across Asia and Africa search for ways to avoid reliance on a handful of American or Chinese firms. The debate is no longer just about privacy but about which governance models will define the next balance of power. Economically, the concentration of data streams creates advantages that stretch far beyond core business lines. Firms able to capture and analyze at scale use predictive power to dominate sectors like health, finance, logistics and education. Data becomes a flywheel that protects incumbents and raises barriers for challengers, narrowing innovation instead of expanding it. For society the consequences are deeply personal. Citizens are treated less as individuals and more as raw material: our clicks, habits and emotions turned into value without clear rights of ownership. Regaining agency will require new architectures of trust built on decentralized identity, transparent exchanges and enforceable digital rights. The real question is whether data remains an instrument of concentration or evolves into a commons that strengthens autonomy. If control stays in a few hands, dependency becomes the default. If governance, culture and innovation align, data could serve as the foundation of a more open and resilient society. The issue is not only who holds the data but whether we choose to remain subjects of its power or become architects of its meaning. #leadership #datasupremacy #digitalsociety #governance #sovereignty #ai #trust

Synthetic #data – artificially generated information that mimics real-world data – can fill gaps, protect privacy and enable the testing of new scenarios. The lines between synthetic and real data are blurring, creating new and significant opportunities, as well as serious risks. Business leaders must prioritize oversight and compliance, building robust traceability and provenance systems to govern synthetic data use. Author: Lauren Woodman CEO, DataKind Arun Sundararajan Harold Price Professor, Entrepreneurship and Technology, NYU Stern School of Business, New York University #AMGFCC25 #CybersecurityAwarenessMonth #DataGovernance https://lnkd.in/gnAHm87V

🚨🇺🇸 Breaking News : Nonprofit Sues 4 Federal Agencies Over AI Use in Trump-Era Policy Enforcement A major lawsuit has been filed and what if Exposure Is Apparent on how the federal government is using AI to shape policy. 📌 The nonprofit Democracy Forward is suing four agencies — OPM, GSA, HUD, and OMB — for failing to respond to FOIA requests about AI tools being used in carrying out the Trump administration’s directives. ⚠️ Their concern: AI may have been applied in analyzing public comments and implementing policies through the Department of Government Efficiency (DOGE), raising questions about transparency, accountability, and civil service protections. Skye Perryman, CEO of Democracy Forward, didn’t hold back: “The public has a right to know the extent to which the administration has used unreliable and unproven AI tools to expand its agenda of undermining regulations that protect people.” This case brings up a bigger question — when AI is used in federal rulemaking and enforcement, where’s the line between innovation and unaccountable automation? The four agencies now have 30 days to respond to the lawsuit. 📌 Question for my network: Should AI play a role in federal policy execution at all, or is this a step too far without clear guardrails? What do you Think? #AI #GovTech #Federal #Transparency #Accountability #GovCon #Technology 📰 Source: Nextgov/FCW — Credit Alexandra Kelley