How AI Is Unearthing Hidden Scientific Knowledge | Sara Beery | TED

Overview

This talk explores how artificial intelligence (AI) can revolutionize ecological research and conservation efforts. With 80% of Earth's species still unknown and extinction rates accelerating, Sara Beery, an AI researcher and ecologist, highlights how AI tools can unlock hidden insights from vast ecological datasets, enabling faster, more efficient scientific discovery and action.

Notable Quotes

- It can feel like we're discovering species just in time to write their obituaries. – Sara Beery, on the urgency of biodiversity conservation.

- The future of conservation is hiding in our ecological databases. – Sara Beery, emphasizing the untapped potential of existing data.

- We stand at a unique point in history. We have both an unprecedented biodiversity crisis and unprecedented tools to address it. – Sara Beery, on the intersection of crisis and opportunity.

🌍 The Biodiversity Crisis

- 80% of Earth's species remain undiscovered, with extinction rates now 100 to 1,000 times higher than historical norms.

- Threats include habitat loss, climate change, invasive species, and natural disasters.

- The interconnectedness of ecosystems means the loss of one species can trigger cascading effects, potentially collapsing entire ecosystems.

📸 Hidden Knowledge in Ecological Data

- Platforms like iNaturalist host millions of ecological observations (e.g., 300 million images), but much of the data's potential remains untapped.

- A single image can reveal species interactions, individual animal identification, vegetation types, and even carbon storage.

- Other databases, such as xeno-canto (bioacoustics) and FathomNet (deep-sea footage), hold similarly vast, underutilized ecological insights.

🤖 AI as a Transformative Tool

- Traditional data analysis methods are too slow to keep pace with the biodiversity crisis. AI can process massive datasets in seconds.

- Beery's team developed Inquire, an AI system that allows scientists to ask natural language questions (e.g., Do birds eat insects in winter?) and receive relevant data without coding or extensive training.

- Inquire has already enabled researchers to analyze bird diets across seasons in just three hours—a task that previously took 1,560 hours.

🔬 Expanding Scientific Discovery

- AI tools like Inquire are flexible, enabling research on diverse topics such as forest regeneration after fires, species mortality in urban vs. rural areas, and climate-driven changes in flowering events.

- Future systems could integrate multiple data types (e.g., images, audio, satellite data) to uncover hidden ecological connections.

🌱 Empowering Global Conservation Efforts

- AI maximizes the value of existing data, helping scientists identify knowledge gaps and prioritize new data collection.

- Public participation is crucial: every photo, sound recording, or observation uploaded to platforms like iNaturalist contributes to the broader understanding of Earth's ecosystems.

- Beery envisions a collaborative future where AI and citizen science drive conservation actions to protect biodiversity before it's too late.