Tuesday, July 7, 2026

4:30 PM - 6:00 PM

Room BE.0.17

---

Details

Participants need to bring their own laptops to this workshop. Modern AI tools are transforming large-scale data analysis workflows in conservation, allowing rapid processing of data at an unprecedented scale and speed. However, many researchers struggle to combine these tools with satellite data, generate reproducible code, and produce reliable scientifically-sound outputs. We propose a 90-minute, hands-on workshop that teaches participants a practical, defensible “vibe coding” workflow, using natural language prompts to generate code, for Earth Observation (EO) analyses using Google Earth Engine, Google Colab and Gemini. After short framing talks (policy need for EO and methodological safeguards), participants will work through an AI-supported EO analysis problem that demonstrates: (1) rapid vibe coding analyses of EO-derived imagery, (2) standardized metadata capture and versioning for reproducibility, (3) where and how vibe coding can fail, and (4) how to identify and fix common vibe coding failures. The session will conclude with an open discussion of the outstanding risks and concerns about AI-enabled programming. Emphasis is on transparency, bias-awareness, and producing results that can inform policy-relevant conservation decisions. Participants will leave with an EO analysis Google Colab notebook, a short checklist for robust vibe coding, and guidance for when and how to integrate vibe coding into research. The session is targeted at researchers with basic Python familiarity; all materials are accessible, documented, and released under an open licence. Attendee participation in discussion, case study choice, and learner support will follow EDIJ principles; the session will provide support using scaffolded participation, building on participants’ existing knowledge and then encouraging independent exploration and learning through a selected case study and small-group work. This practical training bridges technical skills with critical thinking needed to apply AI and EO responsibly in effective biodiversity conservation.

---

Organiser