Can AI Save Nature?
Summary
In this opening keynote at ITB Berlin 2026, Dr. Frauke Fischer — biodiversity expert, founder of Agentur Auf!, and author of the book 'Can AI Save Nature?' — delivers a grounding session on the intersection of biodiversity, ecosystem services, and artificial intelligence, followed by a moderated Q&A with Prof. Dr. Willy Legrand of IU International University of Applied Sciences.
Fischer opens by immediately answering the keynote's title question: 'No — only we could, and we actually should.' The session then builds a rigorous case for why biodiversity matters economically and existentially, before exploring how AI can serve as a powerful observational and predictive tool — but not a substitute for human will and behavioral change.
Fischer explains biodiversity across three levels: genetic diversity (within species), species diversity (between species), and ecosystem diversity (processes within ecosystems). She grounds each level in alarming data: within the last 50 years, 73% of all vertebrate species populations have been lost, along with 80% of all insects — not extinctions of species, but catastrophic population collapse. She reframes rarity as 'the first step toward extinction.' Species extinction is now happening at roughly 1,000 times the natural background rate. To dramatize human dominance: of all mammal biomass on Earth today, humans account for 36%, livestock (cows, pigs) 60%, and all wild mammals combined — from shrews to blue whales — just 4%.
Ecosystem services are quantified with precision: forests provide 75% of all available fresh water; producing a single car requires 150,000 liters of fresh water; 80% of approved drugs and 70% of cancer drugs derive from nature; and the annual economic value of animal pollination alone is €1 trillion. The total annual global monetary value of biodiversity and ecosystem services is described as twice the global GDP.
Fischer then pivots to what AI can do. AI excels at pattern recognition in nature — DNA codes, skin markings, soundscapes, movement signatures — and she provides three concrete case studies. First: whale shark identification. Before AI, only 300 whale shark encounters were recorded over a century. AI now scans social media images and identifies individuals by skin pattern, expanding the known population to 3,000 identified individuals. Movement prediction models also help ship operators avoid collisions — a commercial incentive that aligns with conservation. Second: the 'Internet of Animals,' where sensors on wildlife function as real-time environmental monitors. Goats, sheep, deer, and dogs wearing sensors allowed researchers to predict volcano eruptions more accurately than technical equipment. Southern sea elephants — surfacing every 100 meters to breathe — provide 80% of the data needed to model sea level rise in polar oceans, surpassing buoy-based systems. Bird migration patterns help forecast weather and pandemic spread (e.g., avian flu). Third: soundscape analysis of coral reefs, where a submerged microphone and AI assessment can determine reef health and recovery status without invasive species counts.
In Q&A, Fischer is pressed on whether the tourism industry understands its biodiversity footprint. Her answer: 'No.' She invokes the IPBES (Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services), whose latest report states that businesses must either transform or collapse. She uses the analogy of a car driving at 150 km/h toward a wall — hesitating to brake because slowing down might also benefit competitors. The tragedy of the commons is named as a structural problem: ecosystem services are free, exclude no one, and generate no invoice, making them systematically undervalued.
On citizen science and tourism, Fischer acknowledges a catch-22: tourist presence collecting biodiversity data can also disturb sensitive species. Cheetahs, for instance, are so sensitive to human presence that repeated disturbance collapses the survival rate of their young. The resolution, she argues, is education and communication — helping tourists understand that the lion they visit for one day must endure 365 days of visitors per year.
Key Takeaways
- •AI cannot save nature — only humans can. Fischer answers the keynote's own question immediately: 'The answer is no. Only we could, and the very important thing is that we actually should.'
- •In the last 50 years, 73% of all vertebrate species populations and 80% of all insect populations have been lost — not species extinctions, but population collapses that represent the first step toward extinction.
- •Current species extinction is occurring at approximately 1,000 times the natural background rate. Fischer notes that the dinosaur extinction took at least 33,000 years; today, equivalent destruction would take about 33 years.
- •Human and livestock biomass now dominates Earth's mammals: humans 36%, livestock 60%, all wild mammals combined just 4% — from the smallest shrew to the blue whale.
- •Key ecosystem service statistics: forests supply 75% of global fresh water; manufacturing one car requires 150,000 liters of water; 80% of approved drugs and 70% of cancer drugs originate from nature; pollination services are worth €1 trillion annually; total biodiversity and ecosystem services value is estimated at twice global GDP.
- •AI identified 3,000 individual whale sharks by scanning social media images for skin patterns — up from just 300 documented encounters over 100 years before AI. Movement prediction now helps ship operators avoid collisions, creating a commercial conservation incentive.
- •Southern sea elephants provide 80% of the polar ocean data used to model sea level rise, outperforming buoy systems because they naturally dive to and resurface from the top 100 meters of water — the critical data zone.
- •Researchers predicted volcano eruptions more accurately using AI-analyzed movement data from GPS-collared goats, sheep, deer, and dogs than with traditional technical monitoring equipment.
- •Coral reef health can be assessed by recording and AI-analyzing the soundscape of the reef — a faster and less invasive alternative to manual species counts.
Transcript
So for you, we actually beginning then the morning we're beginning the morning with a question that sounds almost philosophical for some of us but is actually deeply urgent is can artificial intelligence save nature. Um and behind the question is perhaps even a more important uh one and do we actually understand nature and biodiversity well enough in tourism to know what we're asking technology to do for us. uh to answer that I'm really delighted to welcome one of the European leading voice on t...
