As the world increasingly leans on data to power AI innovation, the question of how to make data open, accessible, and equitable becomes more pressing. The Linux Foundation recently published a detailed report, Pathways to Open Data, highlighting insights from a challenge session it hosted at the 2024 World Open Innovation Conference. The session explored barriers and enablers in building open data ecosystems essential for the future of technologies like generative AI and large language models (LLMs).

In this TFIR conversation, Swapnil Bhartiya talks with Anna Hermansen, Researcher and Ecosystem Manager at the Linux Foundation, about the current state and future pathways of open data. From the role of GDPR to rethinking data ownership, Hermansen lays out a practical vision for open data in an AI-powered world.

Q&A: Anna Hermansen, Researcher and Ecosystem Manager at the Linux Foundation

Swapnil Bhartiya: Hi, this is your host, Swapnil Bhartiya. The Linux Foundation recently published Pathways to Open Data, a report summarizing its challenge session at the World Open Innovation Conference held last November. Participants discussed the challenges of opening up data, how organizations use it, and opportunities for greater openness. To discuss this report, we have with us Anna Hermansen, researcher and ecosystem manager at the Linux Foundation. Great to have you on the show.

Anna Hermansen: Thanks for having me! I’m excited to be here.

Swapnil Bhartiya: Before we dive into the report findings, tell us about the genesis of this report and the Linux Foundation’s role at the World Open Innovation Conference.

Anna Hermansen: One of our research advisors, Henry Chesbrough, is deeply involved with the open innovation community and helped us organize a challenge session at the conference in Berkeley, California. We focused on “pathways to open data” because, as AI and generative AI go mainstream, the demand for high-quality, diverse datasets is exploding. This includes not just internal data, but external and third-party data too. As an open source community, we asked: how can we expand from open code to open data? This report captures what we learned.

Swapnil Bhartiya: The term “open” can be ambiguous. How do you define “open data” in this report?

Anna Hermansen: We based our definition on the roots of open science. Open data here means non-personal, non-commercial data that is made publicly available to enable transparency, collaboration, and innovation. Specifically, open data must meet technical and legal requirements that allow anyone to use, reuse, and redistribute it—similar to open source licensing. We also discussed openness in the broader sense—such as access to open standards, development control, cost of use, and collaborative models.

Swapnil Bhartiya: We used to say that “data is the new oil,” but now, it’s probably closer to solar energy. What makes open data more critical today than before?

Anna Hermansen: As industries digitize, data drives innovation across sectors—from marketing and healthcare to finance. Data enables compliance, such as KYC and AML, and accelerates R&D. But data has become so valuable that companies often silo it for competitive advantage or to comply with privacy regulations. With LLMs and AI, the need to train models using diverse, external datasets is critical, yet access remains limited. That tension sparked this conversation about creating a responsible, open data ecosystem.

Swapnil Bhartiya: Aside from legal restrictions like GDPR, what are the other major challenges around open data that participants highlighted?

Anna Hermansen: Several unique barriers emerged. First, data is not like software—it’s malleable, costly to maintain, and comes with complex licensing. Accuracy is paramount, especially with PII. Participants noted that open datasets often lack curation and funding, reducing their value. There’s also a lack of standardization, and high risk aversion from legal teams worried about privacy and IP. The combination of technical, economic, and cultural factors makes data openness more complicated than open source code.

Swapnil Bhartiya: What about the impact of regional regulations—say, GDPR in Europe versus U.S. or Asia? Did anyone suggest adopting models from other regions?

Anna Hermansen: Yes, many participants were European, so GDPR came up a lot. Even though it’s a European regulation, it has global implications due to the internet’s borderless nature. In our broader Linux Foundation research—especially in healthcare—we’ve seen more standardization and collaboration in Europe. The EU seems better positioned to coordinate open data efforts than North America, which still faces fragmentation in policy and execution.

Swapnil Bhartiya: Let’s talk about solutions. What pathways forward were discussed?

Anna Hermansen: One idea was a semi-open platform where competitors share pre-competitive data—like best practices—without exposing customer or proprietary details. We also referenced the Overture Maps Foundation, a Linux Foundation project that offers open and interoperable map data.
Conceptually, we explored rethinking data ownership, proposing that individuals—not corporations—should control their data. That includes usage rights and consent-driven sharing models. Another big idea was treating open data as a value-add, promoting collaboration and innovation while creating positive externalities in sectors like healthcare. Finally, there’s a need for neutral governance to manage datasets responsibly—handling access, security, and maintenance in a structured way.

Swapnil Bhartiya: Anna, thank you so much for joining me today and for sharing these insights. Looking forward to our next conversation.

Anna Hermansen: Thank you for having me. Great to chat with you!