Inspectable AI for Science: A Research Object Approach to Generative AI Governance
Summary: arXiv:2604.11261v1 Announce Type: new
This paper introduces AI as a Research Object (AI-RO), a paradigm for governing the use of generative AI in scientific research. Instead of debating whether AI is an author or merely a tool, we propose treating AI interactions as structured, inspectable components of the research process. Under this view, the legitimacy of an AI-assisted scientific paper depends on how model use is integrated into the workflow, documented, and made accountable.
Abstract
Drawing on Research Object theory and FAIR principles, we propose a framework for recording model configuration, prompts, and outputs through interaction logs and metadata packaging. These properties are particularly consequential in security and privacy (S&P) research, where provenance artifacts must satisfy confidentiality constraints, integrity guarantees, and auditability requirements that generic disclosure practices do not address.
Key Components of AI Governance
To effectively govern generative AI in scientific research, we outline several key components:
- Structured Documentation: Documentation of AI interactions must be systematic and detailed, ensuring that all aspects of AI usage are recorded.
- Controlled Disclosure: The information shared about AI’s role and contributions should be carefully managed to protect sensitive data and ensure compliance with ethical standards.
- Integrity-Preserving Provenance Capture: Provenance records must be created in a way that maintains the integrity of the data and the research process.
Implementation of a Lightweight Writing Pipeline
We implement a lightweight writing pipeline in which a language model synthesizes human-authored structured literature review notes under explicit constraints and produces a verifiable provenance record. This approach not only enhances the efficiency of the research process but also ensures that the contributions of AI are clearly documented and accountable.
Position and Initial Demonstrative Workflow
We present this work as a position supported by an initial demonstrative workflow, arguing that governance of generative AI in science can be implemented through structured documentation, controlled disclosure, and integrity-preserving provenance capture. This foundational approach allows researchers to leverage AI responsibly while maintaining the rigor and integrity of scientific inquiry.
Future Developments for Practical Adoption
Based on this example, we outline and motivate a set of necessary future developments required to make such practices practical and widely adoptable:
- Standardization of Protocols: Developing standardized protocols for AI interactions will facilitate consistency across research disciplines.
- Training and Resources: Providing training and resources for researchers on effective AI governance practices is essential for broad adoption.
- Collaboration Across Disciplines: Encouraging interdisciplinary collaboration will enhance the robustness of AI governance frameworks.
Conclusion
As generative AI becomes increasingly integrated into scientific research, establishing a clear framework for governance is imperative. The AI-RO paradigm offers a structured approach to ensure accountability and integrity in the use of AI, ultimately enhancing the credibility and reliability of scientific findings.