Halting mass extinction
Species are disappearing far faster than the natural rate; stopping it needs habitat protection at huge scale, not a single technology.
Open in the interactive tree →The main drivers are land and sea use change, direct exploitation, climate change, pollution and invasive species. The IPBES assessment of 2019 counted about one million species at risk. The 2022 Kunming-Montreal framework aims to protect 30% of land and sea by 2030.
As of October 2026
The IUCN Red List update of 9 July 2026 covers 175,909 assessed species, of which 49,505 are threatened with extinction, only a fraction of the roughly two million described species. The same update found 62% of the 201 endemic hydrothermal-vent mollusc species at risk from deep-sea mining and confirmed five Australian marsupials extinct. The Kunming-Montreal framework is not legally binding. At COP16 (2024) only 44 of 196 parties had submitted updated national biodiversity plans (51 by May 2025); COP17 in Yerevan (19-30 October 2026) holds the first global stocktake of the framework.
What is missing
- Money: a biodiversity finance gap of about US$700 billion per year
- Slowing the conversion of forest and wetland to farmland, which depends on food and land politics
- Monitoring: most species are unassessed, so trends are unknown
- Enforcement against poaching, overfishing and illegal logging
- Limiting climate change that shifts habitats
Becomes possible once solved
- Intact ecosystems that keep supplying water, pollination and fisheries
- Stable food webs and carbon stores
- Recovery of damaged habitats over decades
Open steps
- Monitoring unassessed species High AI leverageIdentify and count species automatically from camera traps, sound and environmental DNA, since most species have never been assessed.
- Extinction risk for unassessed species High AI leveragePredict Red List risk categories for species without assessments from traits, range and threat data, and say where assessors should look first.
- Spotting threats from space High AI leverageDetect forest clearing, selective logging and illegal fishing early enough for rangers and agencies to act on the alerts.
- Habitat shifts under warming Medium AI leveragePredict where species ranges will move under warming and what protected-area networks must cover, including data-poor species.
Where AI could help
Medium AI leverage. AI greatly widens monitoring of species, extinction risk and threats from space; land use, money and enforcement still decide the outcome.
- Identify species automatically in millions of camera-trap images and sound recordings
- Predict extinction risk for unassessed species so Red List effort goes where it matters
- Detect forest clearing, illegal fishing and poaching hotspots from satellite and sensor streams
- Model how species ranges shift under climate change to place protected areas
Shown so far
- In March 2025, Google released SpeciesNet as open source, a camera-trap classifier trained on over 65 million images that assigns images to more than 2,000 animal, taxon and non-animal labels. source
- In May 2022, machine learning predicted IUCN extinction-risk categories for 4,369 reptile species, 3,286 of them unassessed, with about 90% accuracy for threatened versus not threatened. source
Prerequisites
Unlocks
- Recovering biodiversity2060s?