Data Governance Meets the EU AI Act

Curated on Posted on by Stefaan Verhulst

Article by Axel Schwanke: “..The EU AI Act emphasizes sustainable AI through robust data governance, promoting principles like data minimization, purpose limitation, and data quality to ensure responsible data collection and processing. It mandates measures such as data protection impact assessments and retention policies. Article 10 underscores the importance of effective data management in fostering ethical and sustainable AI development…This article states that high-risk AI systems must be developed using high-quality data sets for training, validation, and testing. These data sets should be managed properly, considering factors like data collection processes, data preparation, potential biases, and data gaps. The data sets should be relevant, representative, error-free, and complete as much as possible. They should also consider the specific context in which the AI system will be used. In some cases, providers may process special categories of personal data to detect and correct biases, but they must follow strict conditions to protect individuals’ rights and freedoms…

However, achieving compliance presents several significant challenges:

  • Ensuring Dataset Quality and Relevance: Organizations must establish robust data and AI platforms to prepare and manage datasets that are error-free, representative, and contextually relevant for their intended use cases. This requires rigorous data preparation and validation processes.
  • Bias and Contextual Sensitivity: Continuous monitoring for biases in data is critical. Organizations must implement corrective actions to address gaps while ensuring compliance with privacy regulations, especially when processing personal data to detect and reduce bias.
  • End-to-End Traceability: A comprehensive data governance framework is essential to track and document data flow from its origin to its final use in AI models. This ensures transparency, accountability, and compliance with regulatory requirements.
  • Evolving Data Requirements: Dynamic applications and changing schemas, particularly in industries like real estate, necessitate ongoing updates to data preparation processes to maintain relevance and accuracy.
  • Secure Data Processing: Compliance demands strict adherence to secure processing practices for personal data, ensuring privacy and security while enabling bias detection and mitigation.

Example: Real Estate Data
Immowelt’s real estate price map, awarded as the top performer in a 2022 test of real estate price maps, exemplifies the challenges of achieving high-quality datasets. The prepared data powers numerous services and applications, including data analysis, price predictions, personalization, recommendations, and market research…(More)”

Why Digital Public Goods, including AI, Should Depend on Open Data

Curated on Posted on by Stefaan Verhulst

Article by Cable Green: “Acknowledging that some data should not be shared (for moral, ethical and/or privacy reasons) and some cannot be shared (for legal or other reasons), Creative Commons (CC) thinks there is value in incentivizing the creation, sharing, and use of open data to advance knowledge production. As open communities continue to imagine, design, and build digital public goods and public infrastructure services for education, science, and culture, these goods and services – whenever possible and appropriate – should produce, share, and/or build upon open data.

Open Data and Digital Public Goods (DPGs)

CC is a member of the Digital Public Goods Alliance (DPGA) and CC’s legal tools have been recognized as digital public goods (DPGs). DPGs are “open-source software, open standards, open data, open AI systems, and open content collections that adhere to privacy and other applicable best practices, do no harm, and are of high relevance for attainment of the United Nations 2030 Sustainable Development Goals (SDGs).” If we want to solve the world’s greatest challenges, governments and other funders will need to invest in, develop, openly license, share, and use DPGs.

Open data is important to DPGs because data is a key driver of economic vitality with demonstrated potential to serve the public good. In the public sector, data informs policy making and public services delivery by helping to channel scarce resources to those most in need; providing the means to hold governments accountable and foster social innovation. In short, data has the potential to improve people’s lives. When data is closed or otherwise unavailable, the public does not accrue these benefits.CC was recently part of a DPGA sub-committee working to preserve the integrity of open data as part of the DPG Standard. This important update to the DPG Standard was introduced to ensure only open datasets and content collections with open licenses are eligible for recognition as DPGs. This new requirement means open data sets and content collections must meet the following criteria to be recognised as a digital public good.

  1. Comprehensive Open Licensing:
    1. The entire data set/content collection must be under an acceptable open licence. Mixed-licensed collections will no longer be accepted.
  2. Accessible and Discoverable:
    1. All data sets and content collection DPGs must be openly licensed and easily accessible from a distinct, single location, such as a unique URL.
  3. Permitted Access Restrictions:
    1. Certain access restrictions – such as logins, registrations, API keys, and throttling – are permitted as long as they do not discriminate against users or restrict usage based on geography or any other factors…(More)”.

The Case for Local and Regional Public Engagement in Governing Artificial Intelligence

Curated on Posted on by Stefaan Verhulst

Article by Stefaan Verhulst and Claudia Chwalisz: “As the Paris AI Action Summit approaches, the world’s attention will once again turn to the urgent questions surrounding how we govern artificial intelligence responsibly. Discussions will inevitably include calls for global coordination and participation, exemplified by several proposals for a Global Citizens’ Assembly on AI. While such initiatives aim to foster inclusivity, the reality is that meaningful deliberation and actionable outcomes often emerge most effectively at the local and regional levels.

Building on earlier reflections in “AI Globalism and AI Localism,” we argue that to govern AI for public benefit, we must prioritize building public engagement capacity closer to the communities where AI systems are deployed. Localized engagement not only ensures relevance to specific cultural, social, and economic contexts but also equips communities with the agency to shape both policy and product development in ways that reflect their needs and values.

While a Global Citizens’ Assembly sounds like a great idea on the surface, there is no public authority with teeth or enforcement mechanisms at that level of governance. The Paris Summit represents an opportunity to rethink existing AI governance frameworks, reorienting them toward an approach that is grounded in lived, local realities and mutually respectful processes of co-creation. Toward that end, we elaborate below on proposals for: local and regional AI assemblies; AI citizens’ assemblies for EU policy; capacity-building programs, and localized data governance models…(More)”.

Good government data requires good statistics officials – but how motivated and competent are they?

Curated on Posted on by Stefaan Verhulst

Worldbank Blog: “Government data is only as reliable as the statistics officials who produce it. Yet, surprisingly little is known about these officials themselves. For decades, they have diligently collected data on others –  such as households and firms – to generate official statistics, from poverty rates to inflation figures. Yet, data about statistics officials themselves is missing. How competent are they at analyzing statistical data? How motivated are they to excel in their roles? Do they uphold integrity when producing official statistics, even in the face of opposing career incentives or political pressures? And what can National Statistical Offices (NSOs) do to cultivate a workforce that is competent, motivated, and ethical?

We surveyed 13,300 statistics officials in 14 countries in Latin America and the Caribbean to find out. Five results stand out. For further insights, consult our Inter-American Development Bank (IDB) report, Making National Statistical Offices Work Better.

1. The competence and management of statistics officials shape the quality of statistical data

Our survey included a short exam assessing basic statistical competencies, such as descriptive statistics and probability. Statistical competence correlates with data quality: NSOs with higher exam scores among employees tend to achieve better results in the World Bank’s Statistical Performance Indicators (r = 0.36).

NSOs with better management practices also have better statistical performance. For instance, NSOs with more robust recruitment and selection processes have better statistical performance (r = 0.62)…(More)”.

Which Health Facilities Have Been Impacted by L.A.-Area Fires? AI May Paint a Clearer Picture

Curated on Posted on by Stefaan Verhulst

Article by Andrew Schroeder: “One of the most important factors for humanitarian responders in these types of large-scale disaster situations is to understand the effects on the formal health system, upon which most people — and vulnerable communities in particular — rely upon in their neighborhoods. Evaluation of the impact of disasters on individual structures, including critical infrastructure such as health facilities, is traditionally a relatively slow and manually arduous process, involving extensive ground truth visitation by teams of assessment professionals.

Speeding up this process without losing accuracy, while potentially improving the safety and efficiency of assessment teams, is among the more important analytical efforts Direct Relief can undertake for response and recovery efforts. Manual assessments can now be effectively paired with AI-based analysis of satellite imagery to do just that…

With the advent of geospatial AI models trained on disaster damage impacts, ground assessment is not the only tool available to response agencies and others seeking to understand how much damage has occurred and the degree to which that damage may affect essential services for communities. The work of the Oregon State University team of experts in remote sensing-based post-disaster damage detection, led by Jamon Van Den Hoek and Corey Scher, was featured in the Financial Times on January 9.

Their modeling, based on Sentinel-1 satellite imagery, identified 21,757 structures overall, of which 11,124 were determined to have some level of damage. The Oregon State model does not distinguish between different levels of damage, and therefore cannot respond to certain types of questions that the manual inspections can respond to, but nevertheless the coverage area and the speed of detection have been much greater…(More)”.

Kickstarting Collaborative, AI-Ready Datasets in the Life Sciences with Government-funded Projects

Curated on Posted on by Stefaan Verhulst

Article by Erika DeBenedictis, Ben Andrew & Pete Kelly: “In the age of Artificial Intelligence (AI), large high-quality datasets are needed to move the field of life science forward. However, the research community lacks strategies to incentivize collaboration on high-quality data acquisition and sharing. The government should fund collaborative roadmapping, certification, collection, and sharing of large, high-quality datasets in life science. In such a system, nonprofit research organizations engage scientific communities to identify key types of data that would be valuable for building predictive models, and define quality control (QC) and open science standards for collection of that data. Projects are designed to develop automated methods for data collection, certify data providers, and facilitate data collection in consultation with researchers throughout various scientific communities. Hosting of the resulting open data is subsidized as well as protected by security measures. This system would provide crucial incentives for the life science community to identify and amass large, high-quality open datasets that will immensely benefit researchers…(More)”.

Announcing SPARROW: A Breakthrough AI Tool to Measure and Protect Earth’s Biodiversity in the Most Remote Places

Curated on Posted on by Stefaan Verhulst

Blog by Juan Lavista Ferres: “The biodiversity of our planet is rapidly declining. We’ve likely reached a tipping point where it is crucial to use every tool at our disposal to help preserve what remains. That’s why I am pleased to announce SPARROW—Solar-Powered Acoustic and Remote Recording Observation Watch, developed by Microsoft’s AI for Good Lab. SPARROW is an AI-powered edge computing solution designed to operate autonomously in the most remote corners of the planet. Solar-powered and equipped with advanced sensors, it collects biodiversity data—from camera traps, acoustic monitors, and other environmental detectors—that are processed using our most advanced PyTorch-based wildlife AI models on low-energy edge GPUs. The resulting critical information is then transmitted via low-Earth orbit satellites directly to the cloud, allowing researchers to access fresh, actionable insights in real time, no matter where they are. 

Think of SPARROW as a network of Earth-bound satellites, quietly observing and reporting on the health of our ecosystems without disrupting them. By leveraging solar energy, these devices can run for a long time, minimizing their footprint and any potential harm to the environment…(More)”.

It Was the Best of Times, It Was the Worst of Times: The Dual Realities of Data Access in the Age of Generative AI

Curated on Posted on by Stefaan Verhulst

Article by Stefaan Verhulst: “It was the best of times, it was the worst of times… It was the spring of hope, it was the winter of despair.” –Charles Dickens, A Tale of Two Cities

Charles Dickens’s famous line captures the contradictions of the present moment in the world of data. On the one hand, data has become central to addressing humanity’s most pressing challenges — climate change, healthcare, economic development, public policy, and scientific discovery. On the other hand, despite the unprecedented quantity of data being generated, significant obstacles remain to accessing and reusing it. As our digital ecosystems evolve, including the rapid advances in artificial intelligence, we find ourselves both on the verge of a golden era of open data and at risk of slipping deeper into a restrictive “data winter.”

A Tale of Two Cities by Charles Dickens (1902)

These two realities are concurrent: the challenges posed by growing restrictions on data reuse, and the countervailing potential brought by advancements in privacy-enhancing technologies (PETs), synthetic data, and data commons approaches. It argues that while current trends toward closed data ecosystems threaten innovation, new technologies and frameworks could lead to a “Fourth Wave of Open Data,” potentially ushering in a new era of data accessibility and collaboration…(More)” (First Published in Industry Data for Society Partnership’s (IDSP) 2024 Year in Review).

The Emergence of National Data Initiatives: Comparing proposals and initiatives in the United Kingdom, Germany, and the United States

Curated on Posted on by Stefaan Verhulst

Article by Stefaan Verhulst and Roshni Singh: “Governments are increasingly recognizing data as a pivotal asset for driving economic growth, enhancing public service delivery, and fostering research and innovation. This recognition has intensified as policymakers acknowledge that data serves as the foundational element of artificial intelligence (AI) and that advancing AI sovereignty necessitates a robust data ecosystem. However, substantial portions of generated data remain inaccessible or underutilized. In response, several nations are initiating or exploring the launch of comprehensive national data strategies designed to consolidate, manage, and utilize data more effectively and at scale. As these initiatives evolve, discernible patterns in their objectives, governance structures, data-sharing mechanisms, and stakeholder engagement frameworks reveal both shared principles and country-specific approaches.

This blog seeks to start some initial research on the emergence of national data initiatives by examining three national data initiatives and exploring their strategic orientations and broader implications. They include:

Impact Inversion

Curated on Posted on by Stefaan Verhulst

Blog by Victor Zhenyi Wang: “The very first project I worked on when I transitioned from commercial data science to development was during the nadir between South Africa’s first two COVID waves. A large international foundation was interested in working with the South African government and a technology non-profit to build an early warning system for COVID. The non-profit operated a WhatsApp based health messaging service that served about 2 million people in South Africa. The platform had run a COVID symptoms questionnaire which the foundation hoped could help the government predict surges in cases.

This kind of data-based “nowcasting” proved a useful tool in a number of other places e.g. some cities in the US. Yet in the context of South Africa, where the National Department of Health was mired in serious capacity constraints, government stakeholders were bearish about the usefulness of such a tool. Nonetheless, since the foundation was interested in funding this project, we went ahead with it anyway. The result was that we pitched this “early warning system” a handful of times to polite public health officials but it was otherwise never used. A classic case of development practitioners rendering problems technical and generating non-solutions that primarily serve the strategic objectives of the funders.

The technology non-profit did however express interest in a different kind of service — what about a language model that helps users answer questions about COVID? The non-profit’s WhatsApp messaging service is menu-based and they thought that a natural language interface could provide a better experience for users by letting them engage with health content on their own terms. Since we had ample funding from the foundation for the early warning system, we decided to pursue the chatbot project.

The project has now spanned to multiple other services run by the same non-profit, including the largest digital health service in South Africa. The project has won multiple grants and partnerships, including with Google, and has spun out into its own open source library. In many ways, in terms of sheer number of lives affected, this is the most impactful project I have had the privilege of supporting in my career in development, and I am deeply grateful to have been part of the team involved bringing it into existence.

Yet the truth is, the “impact” of this class of interventions remain unclear. Even though a large randomized controlled trial was done to assess the impact of the WhatsApp service, such an evaluation only captures the performance of the service on outcome variables determined by the non-profit, not on whether these outcomes are appropriate. It certainly does not tell us whether the service was the best means available to achieve the ultimate goal of improving the lives of those in communities underserved by health services.

This project, and many others that I have worked on as a data scientist in development, uses an implicit framework for impact which I describe as the design-to-impact pipeline. A technology is designed and developed, then its impact is assessed on the world. There is a strong emphasis to reform, to improve the design, development, and deployment of development technologies. Development practitioners have a broad range of techniques to make sure that the process of creation is ethical and responsible — in some sense, legitimate. With the broad adoption of data-based methods of program evaluation, e.g. randomized control trials, we might even make knowledge claims that an intervention truly ought to bring certain benefits to communities in which the intervention is placed. This view imagines that technologies, once this process is completed, is simply unleashed onto the world, and its impact is simply what was assessed ex ante. An industry of monitoring and evaluation surrounds its subsequent deployment; the relative success of interventions depends on the performance of benchmark indicators…(More)”.