I Built Gemini-Scientibots to Fix What's Broken in Science.
By Marie-Soleil Seshat Landry
For as long as I can remember, I've been driven by two passions: the intricate logic of code and the boundless curiosity of the scientific method. My namesake, Seshat, was the ancient Egyptian goddess of wisdom and writing. I've always felt a pull to not just discover knowledge, but to improve how we record and share it.
I've lived the life of a researcher. I know the thrill of a breakthrough and the profound satisfaction of seeing your data support your hypothesis. But I also know the grind. The countless hours spent formatting citations, the late nights wrestling with statistical software, and the daunting blank page of a new manuscript. I've seen brilliant minds—colleagues and friends right here in our community—bogged down by tedious, administrative work that stifles the very creativity it's meant to support.
That friction, that waste of human potential, is what drove me to create Gemini-Scientibots. I wanted to build the tool I wish I'd had—an AI partner that handles the grunt work, so scientists can get back to the business of science.
Phase 1: Automating the Entire Pre-Lab Process 📝
The journey from a simple observation to a robust experimental plan is where many projects lose momentum. I designed Gemini-Scientibots to turn this multi-week process into a matter of hours.
It starts with your idea. You provide the AI with a general observation. From there, it initiates a collaborative dialogue to:
- Refine Your Idea into multiple, focused, and testable research questions.
- Conduct a Deep Literature Search to synthesize existing knowledge and, crucially, identify a clear gap for your research to fill.
- Formulate a Falsifiable Hypothesis (H_a) and its corresponding null (H_0), giving your experiment statistical and intellectual rigor from day one.
- Generate a Lab-Ready SOP: This is the cornerstone. The AI produces a complete Standard Operating Procedure in LaTeX, detailing everything from materials and methods to the pre-defined statistical analysis plan.
The end product of this phase is a "pre-experimental preprint." The Introduction and Methods sections of your paper are already written and impeccably formatted. You have a clear roadmap before you even pick up a pipette.
The Irreplaceable Scientist: You 🧑🔬
Let me be perfectly clear: I did not build Gemini-Scientibots to replace researchers. I built it to empower them. The AI's role stops at the lab door.
The physical execution of the experiment—the skilled hands, the watchful eye that catches a subtle colour change, the intuition to note a deviation from the plan—that is, and must always be, the domain of the human scientist. This is where discovery truly happens.
Phase 2: From Raw Data to a Complete Draft 📄
Once your experiment is done, the partnership with the AI resumes. You provide it with your raw data. From there, Gemini-Scientibots executes the post-lab workflow:
- Automated Analysis & Visualization: It processes your data according to the plan laid out in the SOP, runs the specified statistical tests, and generates publication-quality figures and tables. No more struggling with graphing software.
- Full Manuscript Assembly: The AI then integrates everything—the intro and methods from phase one, and the new results and figures—into a single, cohesive manuscript in the standard IMRaD format. The Results section is written objectively, and the bibliography is perfectly formatted.
My Guiding Principle: Elevating Human Intellect
The most important part of any scientific paper is the interpretation—the meaning of the results. This is not a task for an algorithm.
That's why I programmed Gemini-Scientibots to generate the Discussion and Conclusion sections with guided placeholders. It presents the completed results and then asks you, the expert, the critical questions: Do these findings support your hypothesis? What are the broader implications? What are the limitations, and what should be done next?
This structure ensures the final product is a powerful synthesis of machine efficiency and irreplaceable human intellect.
My vision for Gemini-Scientibots is simple: to accelerate the pace of discovery by freeing scientists from drudgery. It's a tool designed to let you spend less time formatting and more time thinking. Here in Moncton, and in labs around the world, there are brilliant ideas waiting to be explored. My hope is that this tool helps bring them to light, faster than ever before.
To see this process in action and understand its potential for your own work, I invite you to explore an interactive demonstration of the Gemini-Scientibots workflow at the following link:
https://gemini.google.com/gem/1ug682BdcRkOcWfCQmLCJsSSlxyPYy6Pn?usp=sharing
Let's build the future of research together.
Marie-Soleil Seshat Landry Creator, Gemini-Scientibots
Comments
Post a Comment