AI summary by TwinMind
Tanya Berger-Wolf: How your nature photos can help protect wild animals | TED
TED
9m 29s
The Biodiversity Crisis
- Sudan, last male northern white rhino, died in 2018 leaving only 2 females
- Currently in sixth mass extinction/biodiversity crisis
- Lack scientific/technological solutions to track biodiversity loss rate
- IUCN Red List tracks 130,000 species out of millions existing
- Majority have conservation status as "data deficient"
- Population trends marked as "unknown"
- Includes iconic species: killer whales, polar bears
- Critical data gaps prevent effective conservation policy decisions
- Unknown: African elephant populations, poaching rates, whale migration patterns, juvenile turtle survival rates
Traditional Wildlife Tracking Limitations
- Insufficient scientists to track all animals globally
- Limited satellite tags and collars available
- Tagging process requires:
- Animal capture and tranquilization
- Veterinary monitoring of vital signs
- Risk of collar snagging or tag infection
- Dangerous to animals despite precautions
AI-Powered Solution: Wildbook Platform
- Images most abundant data source available
- Sources: scientists, field assistants, camera traps, drones, tourists
- Millions of images from safaris, whale watching tours
- Wildbook uses modern AI, machine learning, computer vision
- Capabilities:
- Find images containing animals
- Locate animals in pictures (including baby elephant behind mother)
- Identify species and individual animals
- Recognize unique markings as "body prints"
- Stripes, spots, wrinkles, notches
- Whale fluke shapes, dolphin dorsal fins
- Track animals using photos instead of tags/collars
- Count populations
- Map social networks and animal relationships
Flukebook Case Study
- Wildbook for whales and dolphins
- Features "Pinchy" - most sighted animal with 600+ sightings around Dominica
- Contains 1+ million images of 46,000+ identified individuals
- Provides foundation for science and conservation research
Social Media AI Agent
- Scours publicly posted images and videos
- Identifies animals via machine learning
- Adds data to appropriate Wildbook pages
- Comments on social media posts with identification results
- Example: "At 2 minutes 46 seconds found this whale shark, here's what we know"
- Engages public: "How can I help?"
- Converts vacation videos into scientific data
Whale Shark Wildbook Results
- 12,000+ individual whale sharks tracked
- Data sources:
- 9,000 citizen scientists
- 200+ conservation/science projects
- 1 intelligent social media agent
- Foundation for IUCN Red List entry
- Conservation status changed: vulnerable → endangered
- Population trend: stable → decreasing
- Improved data enables better policy decisions and resource allocation
Grevy's Zebra Census Innovation
- First full species census using AI technology
- January 2016: 2-day photography event in Kenya
- Participants: rangers, school kids, tourists with telephoto cameras
- 40,000+ images collected
- Machine learning identified all individual animals
- Most accurate species count achieved
- Kenya Wildlife Service adopted method for biennial tracking
- "Great Grevy's Rally" repeated in 2018 (1,000+ people), 2020
- Data became basis for:
- IUCN Red List entry
- Conservation policy
- Endangered species management plan
Platform Scale and Impact
- 53 species covered across marine and terrestrial environments
- AI democratizes science by connecting individual cameras into global biodiversity view
- Creates conservation policy and science at large scale, high resolution
- Wild Me non-profit organization hosts Wildbook platform
- Thousands of global contributors:
- Photograph animals
- Annotate images for AI training
- Develop technology
- Field conservation work
