SIPS 2026 DC

In the 2025 In-person SIPS, we introduced the concept of "Radical Transparency" (RT), a practice consisting of making public the whole process of developing a research process, instead of just its outcomes.
To carry out research in this manner, we may leverage existing technologies (e.g., version control systems).

However, there are still many unknowns yet to be uncovered.
In this hackathon, we will focus on trying to make progress towards the following objectives:

• To define the scope and constraints of the RT approach.

• To determine the ability of existing technologies to meet the requirements of RT.

• To uncover the risks and unintended consequences of RT, and propose safeguards to mitigate them.

We will split into groups to draft ideas on these objectives.
At the end of the session, we will gather the results, discuss them, and decide on the next steps towards a publication.

The movement toward open science is driven by the idea of radical transparency. Frameworks such as the OSF enable researchers to share data, materials, and ideas openly, fostering collaboration, efficiency, and reproducibility. At the same time, the rapid rise of AI presents both opportunities and challenges for this vision. Open datasets can accelerate discovery, support new forms of analysis, and power AI-driven tools for scientific progress. However, these same resources may be used to train models without researchers’ knowledge or consent, while preprint repositories may increasingly host rapidly generated and unevenly vetted work.

These developments prompt an important conversation about the future of open science. How can we preserve openness while addressing emerging risks? What norms, infrastructures, or safeguards can ensure transparency supports responsible research? Should we rethink openness or adapt it to a changing technological landscape? This raises a key question: what is the future of open science in an AI-enabled world?

Registered Reports were created to improve openness and replicability in research. While they have gained in popularity over the past decade, they remain difficult to systematically identify and discover, hindering meta-science efforts to evaluate them at scale. To address this gap, we are developing a comprehensive living database of published Registered Reports. Since our 2025 hackathon, initial work focused on building a Zotero database, developing machine learning mechanisms for identifying and collecting Registered Reports, and defining processing terms/information to include in the database. Our hackathon will build directly on this foundation. Participants will collaborate to improve data collection pipelines, improve tagging and metadata standards, expand database coverage across journals and disciplines, and address sustainability, versioning, and long-term maintenance. The goal is to develop a robust, scalable, community-driven resource that enables reliable discovery and reuse of Registered Reports for researchers, meta-researchers, and the broader scientific community.