Free open-source tool · No install · No account

Declare how humans and AI interacted in your knowledge creation process

sIfA gives researchers a structured, standardised way to record AI interaction across every stage of a research project — anchored in the Contributor Role Taxonomy (CRediT) that journals already understand. If CRediT does not apply to your processes, you can use your own process and develop your own taxonomy.

↓ Download online version Requires internet. Loads faster. Recommended for most users.

↓ Download offline version No internet needed after download. For fieldwork or low-connectivity settings.

One HTML file · Open it in Chrome, Edge, Safari, or Firefox · Nothing leaves your computer

What sIfA produces

A brain-shaped figure with one axis for all recorded contributor roles (CRediT or your own taxonomy). The orange field is human contribution; the purple shape is AI interaction; sparkles mark the extent of AI interaction on each axis. Below example shows the sIfA for the sIfA tool.

Example sIfA figure showing AI contribution across CRediT roles

What sIfA is

AI interactions in knowledge creation are expanding faster than conventions for declaring it can keep up. Free-text disclosures are inconsistent and difficult to compare. sIfA — Statement of Intellectual Fellowship and Accountability — gives authors and knowledge creators a shared, structured format for disclosure.

With the sIfA tool you map your AI interaction onto the 14 CRediT contributor roles — the standard that most major journals already use to record who did what. For each role, you record for what type of work AI was used and which AI tool, how extensively (on a simple three point scale: not at all, little, a lot), and a short audit-trail note. The result is a table and a distinctive figure you can attach to any publication, report, or institutional disclosure form. The sIfA figure reflects a commitment to reflection and transparency when it comes to AI in knowledge creation; the table acts as a detailed track record that supports learning for all.

sIfA means ‘the quality of something’ in Arabic and Swahili but also stands for Statement of Intellectual Fellowship and Accountability. The use of the word fellowship to describe AI interaction reflects what advanced work with AI can be: it can be a friendly, supportive and challenging interaction, if used responsibly, reflectively and transparently. "Accountability" keeps responsibility for conducting the work and managing the interaction with AI with the human author, who is making the disclosure deliberately and transparently.

How it works

Download and open

Click the download button above. Save the HTML file anywhere on your computer. Double-click it to open in your browser. No install, no account, no internet required after download. Adjustable to your needs

Fill in your disclosure, ideally from the start of a project to keep an up-to-date record along each step

If using the CRediT framework, work through the 14 CRediT roles. For each role, record who contributed, which AI tool was used (if any), how extensively, and a brief note. Or use the built-in AI agent — it interviews you role by role in plain language. If CRediT does not work for you: enter or build your own taxonomy.

Export and attach

Export your completed sIfA as an SVG figure, a table, or a CSV. Attach it as a supplementary file to your paper, report, or funder submission. Save your working file to return to it later.

Who is it for

Researchers, teachers, creators, students — anyone who interacts with AI in producing an output

Anyone who interacts with AI at any point in a project — for analysis, writing, coding, literature search, brainstorming, project management — and seeks to declare it clearly and consistently.

Research teams (or other teams creating outputs)

Teams with multiple contributors across multiple roles. sIfA records who did what, including AI, in one shared document.

Learners and teachers: students and lecturers.

Learners and teachers seeking to learn and teach processes of knowledge creation and where AI can be helpful or where it hinders learning.

Journal editors

Editors seeking a structured, comparable disclosure format that goes beyond free-text "we used ChatGPT" statements.

Research offices and funders

Institutions building AI disclosure requirements into grant conditions or ethics review processes.

Nothing leaves your computer

sIfA runs entirely in your browser. There is no server, no analytics, no account.

Your contributor names, AI tools used, and audit notes are never uploaded anywhere
Autosave keeps your work in your browser's local storage, on your machine only
Closing the tab does not delete your work — reopen the file to continue
Your exported files are yours to keep, share, or attach as supplementary material. You can even track your AI interaction across multiple projects in your sIfA portfolio.

How to cite sIfA

Cite the tool:

Schomerus, Mareike. sIfA Tool (Tool to create a Statement of Intellectual Fellowship and Accountability). Version 1.3. 2026. Busara. https://doi.org/10.5281/zenodo.20285993

Cite the paper:

Saleh, Engy and Mareike Schomerus. The sIfA tool for a Statement of Intellectual Fellowship and Accountability: An invitation to reflect on how humans and AI interact in producing knowledge. (Groundwork 32 Thought Piece). Nairobi: Busara. doi: 10.62372/IUPB2137. Read the paper

A machine-readable citation file (CITATION.cff) is included in the repository and powers the "Cite this repository" button on GitHub. The CRediT taxonomy itself is maintained by NISO — please cite it separately in extended documentation.