As AI and emerging technologies continue their development, data stewardship is more essential than ever. The risks and opportunities presented by this technology underscore the need for responsible human oversight on policy and societal challenges and the development of quality data via operational data collaborative infrastructures. Effective data stewardship can empower individuals while facilitating the responsible use of new technologies. 

On March 25, 2026,  Stefaan Verhulst hosted the second edition of the “Data Stewardship Trends to Watch” series for our growing data stewards alumni community (200+). The session focused on data governance innovations and looked at recent trends in the data stewardship environment. 

Check out a recording of the full conversation here. The following resources were discussed during the session:

Fourth Wave of Open Data / Model Context Protocol (MCP) and Open Data 

• A Fourth Wave of Open Data? Exploring the Spectrum of Scenarios for Open Data and Generative AI Discussion followed on how AI could democratize data access through conversational interfaces. Participants also discussed Model Context Protocol (MCP) emerging as a key tool and proposed it be used as digital public infrastructure (once public authorities embed it) and more. 

• AI agents are coming for government. How one big city is letting them in that builds the case on how the city of Boston built their own MCP and is experimenting with its use for public services. They also provide a collection of MCP servers designed for LLM and Agentic accessibility that can be accessed on Github. 

• Civic AI Tools is an open-source project that connects AI assistants to government open data. It uses MCP to give AI models structured access to real datasets from cities like New York, Chicago, San Francisco, Seattle, and Los Angeles - in order to ask plain-language questions and receive output grounded in actual public data. 

• Another similar example is the use of AI agents for policy analysis like the example from DatHere on Github providing a detailed NYC housing policy analysis. 

Importance of Context 

In line with these developments, we are also witnessing a shift from prompt engineering to context engineering. 

• When AI is Fluent in Data but Illiterate in Context builds the case on the importance of context when using AI based on a case in Congo. Context literacy is important because if not taken into account it can lead to: 

• Proxy fragility: Formal-economy indicators miss informal activity, leading to silent misrepresentation.

• Interpretive drift: Models impose external categories that distort local realities.

• Data absence bias: Lack of data is mistaken for lack of activity.

StatGPT

• Amid these developments, official data remains hard to find, interpret, and use, and current GenAI tools often produce plausible but incorrect results, risking trust. In this regard, the International Monetary Fund has pushed the StatGPT report. 

Tabular Foundational Models

 

Tabular foundation models are large-scale pre-trained neural networks that enable "in-context learning" on structured data, such as spreadsheets and SQL databases, eliminating the need for retraining or heavy hyperparameter tuning for new tasks. Select readings and tools are: 

• Accurate predictions on small data with a tabular foundation model

• Supply Chain Greenhouse Gas Emission Factors v1.3 by NAICS-6

• Tabular Foundational Model Tool developed by Prior

• Foundational Model for Enterprise developed by Fundamental Tech 

Data Winter

Despite AI democratization potential, we still face continued decline in data accessibility. This is proven by select evidence below: 

• The State of Open Humanitarian Data 2026 report

• Digital Government Index and Open, Useful and Re-usable Data Index

Data Extraction

Another key concern amidst these developments are data extraction practices addressed by different open data and media communities. 

• Mistral CEO: AI companies should pay a content levy in Europe

• America First, Africa Last? Health data deals and the new scramble for pathogens

Policy Developments

As this is an ever-evolving field, innovations in policies addressing such developments are critical. Some policy developments can be accessed below: 

• Guidelines on promoting and harnessing data access as a tool for advancing human rights and sustainable development in the digital age published by The African Commission on Human and Peoples’ Rights. 

• Article 40(4-11) DSA: Guidance for Applicants published by Coimisiún na Meán. 

Finally, we wrapped the session by highlighting some of our internal updates and developments which can be accessed below: 

We have recently launched our new dedicated website regarding our data stewardship efforts that can be accessed via https://datastewards.net/, a new course is live on Open Data as a Foundation for AI Innovation and we are convening two online dialogues on media collective bargaining power vis-a-vis GenAI.  

In addition, select recent publications include: 

• What is Data Governance: 30 Questions and Answers

• Data Governance Innovations: Emerging practices and trends across the data life cycle

• Policy Developments in Data Access Policy

• Selected Readings on Indigenous Data Governance: 2026 Update

Interested to learn more about data stewardship trends? Sign up for the Data Stewards Network mailing at this link. Or contact us directly at contact@datastewards.net to learn more.