Plants and fungi have been dying off quietly for years—partly because we haven’t been looking in the right places, with the right tools, or frankly, with enough people.
Now the Royal Botanic Gardens, Kew—Britain’s botanical heavyweight—says the era of flying blind is ending. On June 16, 2026, Kew released the sixth edition of its State of the World’s Plants and Fungi report, built with input from more than 400 scientists across 40 countries. The headline isn’t “AI will save nature.” It’s sharper than that: digital tools and machine learning are starting to expose extinctions and collapses that science never properly documented in the first place.
That’s both encouraging and a little damning. Because the crisis isn’t only what’s disappearing—it’s what we never bothered to measure.
Kew’s big message: the biodiversity crisis hides in the blank spots
Kew’s report puts the “holes” in our knowledge front and center. The argument is simple: conservation keeps chasing the species we already know and love—charismatic, well-studied, well-funded—while huge swaths of plant and fungal life remain undercounted, under-monitored, and under-protected.
Ten years after the first report in this series, Kew says the toolbox has changed scale. Digitized museum and herbarium collections, giant databases, satellite and remote-sensing data, on-the-ground sensors, and AI systems that can sort mountains of images and measurements—together they’re shifting the job from “describe what exists” to “find where we’re ignorant, fast.”
And that shift matters politically. If your conservation priorities are based only on what’s already famous, you’ll keep missing the quiet collapses—especially among plants and fungi, which have spent decades as the neglected middle children of biodiversity policy.
AI can watch everything—and still tell you nothing if the data stinks
Where AI is actually taking root in conservation is surveillance. Not the sci-fi kind—the boring, useful kind. Camera traps, acoustic monitors, sensor networks, satellite feeds: they generate more photos, audio, and readings than any ranger team can process by hand. Algorithms do the triage—flagging patterns, spotting anomalies, prioritizing what humans should check.
WWF Belgium’s analysis (“L’IA au cœur de la conservation: entre illusion et protection”) points to the usual suspects: camera traps, biodiversity monitoring, and reducing human–wildlife conflict.
But here’s the trap: dashboards and heat maps can create a false sense of control. A slick interface can make it look like nature is neatly managed—when the system might be trained on thin data, biased coverage, sloppy protocols, or too little human verification. A model that works on one pilot site can faceplant somewhere else with different terrain, species, weather, or human activity.
And even when the detection is solid, AI doesn’t make the hard calls. Conservation is still about decisions people fight over: closing a trail, restricting access, moving an activity, restoring habitat, paying for enforcement. Sensors can scream “problem.” They can’t pass a policy, hire staff, or win a public meeting.
Spain’s test case: using sensors and AI to manage tourism pressure
One place trying to make this practical is Spain, where the International Union for Conservation of Nature (IUCN) describes pilot projects using AI, sensors, and digital tools in parks to protect raptors, bats, and high-mountain wetlands.
The IUCN points to work in Catalonia involving Bonelli’s eagle (a threatened raptor Americans won’t know, but think “rare hawk with expensive real estate needs”) and bats, plus monitoring of high-altitude wetlands in Sierra Nevada.
The enemy here isn’t mysterious: biodiversity loss colliding with booming tourism. The IUCN says conservationists struggled for years to track Bonelli’s eagle and protect nesting areas while visitor traffic surged. The tech is meant to put numbers on what park staff have long suspected—when and where human presence starts pushing ecosystems past their limit.
Arnau Teixedor, a program officer at the IUCN Centre for Mediterranean Cooperation and the project coordinator, summed up the vibe in a line that should be printed on every grant proposal: “It showed us that technology could be an ally, not a distraction, for protecting wildlife.”
That’s the point. Less gadget worship. More operational clarity: Where are the sensitive zones? When does foot traffic become a problem? Which weeks are critical for nesting or breeding? The pressure doesn’t vanish—but it becomes measurable, arguable, and defensible when managers have to justify restrictions.
Science has its own AI problem: speed is great until the literature turns to sludge
This isn’t just about parks and fieldwork. It’s also about how knowledge gets made—and how easily it can get polluted.
Agence Science-Presse, looking ahead to 2026, describes a split reality: AI can boost research productivity, but it can also bury scientists under an avalanche of low-quality output—what it bluntly calls “bouillie,” basically academic mush.
Science-Presse notes that “virtual scientists” can test thousands of configurations quickly, and that generative AI tools have been under real-world testing in labs since 2024. It also flags that 2026 should bring clearer evidence about the harmful impact of a flood of flimsy studies on the scientific workforce.
For conservation, that’s not an inside-baseball concern. Management decisions lean on studies, inventories, and models. If the pipeline fills with shaky results, “faster” turns into “wrong faster,” and scarce money gets steered toward the wrong places. Quality control stops being academic etiquette and becomes field operations: ground-truthing, transparent methods, and the ability to tell a real warning from algorithmic noise.
Finding new species faster sounds great—if the basics are done right
AI boosters also pitch it as a way to speed up species discovery and ecosystem monitoring—helping with identification, classification, and pattern detection. McGill University has described AI as an underused tool for biodiversity protection, and the logic tracks with Kew’s thesis: use automation to spotlight the blind spots.
But the conditions are strict. You need solid reference data, consistent protocols, and results that can be explained—not just spat out. Otherwise, AI doesn’t “discover” anything. It amplifies error with confidence.
Kew’s framing is the most honest version of the pitch: technology should tell humans where to go, what to look at, and where urgency is highest. It doesn’t replace botanists, mycologists, field inventories, or the unglamorous grind of habitat protection and restoration.
What 2026 looks like, if Kew is right, is a reorganization of conservation work: less wandering in the dark, more targeted verification—and faster action where multiple signals line up.
Sources
Royal Botanic Gardens, Kew: State of the World’s Plants and Fungi (6th edition, June 16, 2026); WWF Belgium, “L’IA au cœur de la conservation: entre illusion et protection”; IUCN (Spain projects on raptors, bats, and mountain wetlands); McGill University (AI and biodiversity protection); Agence Science-Presse (“5 choses à surveiller sur l’IA en 2026”).
