WHO WE ARE
Garden Research is an independent research practice studying what happens to knowledge when AI enters the picture.
The question most AI discourse skips: what happens to knowledge itself when the systems producing it are opaque, probabilistic, and optimized for plausibility rather than truth? AI doesn't just speed up research, it changes what counts as evidence, how claims get validated, and who gets to participate in inquiry.
We investigate this through empirical research, foresight projects, and by building frameworks and tools for rigorous AI-mediated research.
Garden is self-funded, independent, and not affiliated with any university, corporation, or think tank. That's by design.
The question most AI discourse skips: what happens to knowledge itself when the systems producing it are opaque, probabilistic, and optimized for plausibility rather than truth? AI doesn't just speed up research, it changes what counts as evidence, how claims get validated, and who gets to participate in inquiry.
We investigate this through empirical research, foresight projects, and by building frameworks and tools for rigorous AI-mediated research.
Garden is self-funded, independent, and not affiliated with any university, corporation, or think tank. That's by design.
- ANGELINA ZAITSEVAFounderM.Sc. in Computer Science, HTW Berlin. B.Sc. in IT Management, GSOM SPbU
Previously Managing Director at Bolder, a strategic branding agency working with deep-tech startups — where she spent years watching how knowledge fails to move between technical, institutional, and public contexts. That problem became the research.
Current academic focus: validation layers for LLM-based research systems. How to trace, assess, and verify claims generated by language models before they get treated as knowledge.
CURRENT WORK
Garden Intelligence Engine
A framework for validating the epistemic quality of LLM outputs in research contexts. Language models are already embedded in knowledge production, from literature review to data synthesis, but no established methodology exists for assessing whether their outputs meet the standards of rigorous inquiry. Intelligence Engine proposes a layered validation approach.
Herbarium
A signal-tracking project mapping how AI enters scientific practice in real time. Each edition curates recent studies, tools, and experiments. Published regularly on gardenresearch.eu and Substack.
→ Explore the project
Unspoken Futures 2026
Systematic analysis of 184 global trend reports. The project extracted 420+ claims about the future and mapped institutional consensus and contestation through original visualizations.
→ Explore the project
A framework for validating the epistemic quality of LLM outputs in research contexts. Language models are already embedded in knowledge production, from literature review to data synthesis, but no established methodology exists for assessing whether their outputs meet the standards of rigorous inquiry. Intelligence Engine proposes a layered validation approach.
Herbarium
A signal-tracking project mapping how AI enters scientific practice in real time. Each edition curates recent studies, tools, and experiments. Published regularly on gardenresearch.eu and Substack.
→ Explore the project
Unspoken Futures 2026
Systematic analysis of 184 global trend reports. The project extracted 420+ claims about the future and mapped institutional consensus and contestation through original visualizations.
→ Explore the project
WHAT'S NEXT
Garden is moving from independent publishing to a structured research program. The priorities:
Doctoral research in computer science. Expanding the Intelligence Engine into a full methodological framework for epistemic validation of AI-generated knowledge.
Building collaborations with researchers and institutions working on the same problem: how do we maintain epistemic integrity when AI mediates how we know things?
Scaling Herbarium into a broader observatory for AI in science.
Garden Research is open to research partnerships and institutional collaborations. Let's get in touch.
Doctoral research in computer science. Expanding the Intelligence Engine into a full methodological framework for epistemic validation of AI-generated knowledge.
Building collaborations with researchers and institutions working on the same problem: how do we maintain epistemic integrity when AI mediates how we know things?
Scaling Herbarium into a broader observatory for AI in science.
Garden Research is open to research partnerships and institutional collaborations. Let's get in touch.
