Artificial intelligence is rapidly changing what scientists can see in nature—and what they’ve been missing. On June 16, 2026, the Royal Botanic Gardens, Kew, will publish the sixth edition of its global assessment, State of the World’s Plants and Fungi, drawing on work from more than 400 scientists across 40 countries.
The report’s central argument is blunt: part of the biodiversity crisis is happening in the blind spots of science itself. Kew says digital tools—ranging from digitized museum and herbarium collections to remote sensing, field sensors and AI—are now making it possible to detect declines and extinctions that previously went undocumented.
But Kew also frames the technology as a guide, not a replacement for fieldwork: it can help conservationists decide where to go, what to look for, and where action is most urgent. That same logic is showing up in on-the-ground pilots, including projects described by the International Union for Conservation of Nature (IUCN) in Spain, where sensors and AI are being tested to manage human pressure on sensitive sites.
Kew’s June 16, 2026 report puts “holes” in global knowledge at the center
Kew’s starting point is a methodological warning: the living world is also being lost in places scientists aren’t measuring well. The sixth State of the World’s Plants and Fungi, published June 16, 2026, compiles expertise from more than 400 scientists in 40 countries and argues that digital tools can expose missing data—and steer action toward the areas of greatest urgency, according to Kew.
The report describes a shift in scale since the first edition a decade ago. Databases, collection digitization, remote sensing, field sensors and AI are moving the center of gravity of the work: not just describing biodiversity, but quickly identifying where knowledge is thin—and where inaction carries the highest cost.
That approach has direct implications for conservation policy. Priorities, Kew suggests, can’t be driven only by iconic or well-studied species; they also need to account for poorly documented groups. Plants and fungi—long less monitored than some animals—are presented as areas where technology can reveal quiet declines. Kew calls tech an “ally,” but only if it serves a strategy rather than a headline-grabbing demo.
Cameras, sensors and algorithms are transforming monitoring—not the hard choices on the ground
AI’s most practical entry point into conservation has been monitoring. Automated devices generate volumes of images, audio and measurements that teams can’t realistically process by hand. Algorithms then help sort, detect and prioritize signals. Recent analyses of AI in conservation highlight recurring use cases including camera traps, biodiversity monitoring and preventing conflict between people and wildlife, according to “L’IA au cœur de la conservation: entre illusion et protection.”
The risk, those analyses argue, is that AI can create an illusion of control. A dashboard of maps, alerts and trend lines can make it feel like nature is “managed,” even when results depend heavily on what data gets collected, where coverage exists, what protocols are used, and how much human time goes into verification. The same analyses stress a key boundary between protection and illusion: a system that performs well at a pilot site doesn’t automatically transfer to other environments, according to “L’IA au cœur de la conservation: entre illusion et protection.”
And conservation isn’t just detection. It requires decisions—closing a trail, shifting an activity, increasing enforcement, restoring habitat. Technology can speed diagnosis, but it can’t resolve the social and political tradeoffs. Digital tools can raise the alarm; they don’t create the capacity to act on their own. That gap—between measurement power and decision speed—becomes more consequential as these systems spread.
In Spain, IUCN tests AI to manage tourism pressure on sensitive habitats
The promise of tech becomes more tangible in localized projects. In Spain, the IUCN describes using artificial intelligence, sensors and digital tools in parks to protect raptors, bats and high-mountain wetlands. The IUCN article cites work in Catalonia focused on Bonelli’s eagle and bats, and in Sierra Nevada to monitor high-altitude wetland areas.
At the core is managing a double pressure: biodiversity loss alongside fast-growing tourism. The IUCN says conservationists struggled for years to track Bonelli’s eagle and protect nesting areas amid an influx of visitors. Digital tools, the organization argues, can help quantify the impact of human activity on ecosystems and support park management decisions.
One quote captures the pilot project’s mindset: “It showed us that technology could be an ally, not a distraction, in protecting wildlife,” said Arnau Teixedor, a program officer at the IUCN Centre for Mediterranean Cooperation and the project’s coordinator, according to the IUCN. The line also signals what managers say they want: fewer gadgets, more operational tools that connect observation to decision-making.
In practice, this approach aims to reduce uncertainty when tradeoffs are unavoidable. Where are the most sensitive zones? When does visitor traffic become a problem? Which periods are most critical? Sensors and AI don’t remove the pressure, but they can help document it—making it easier to challenge or defend management measures in public debate and to fine-tune how sites are managed.
Research faces a second AI battle in 2026: speed, “mush,” and quality control
The AI question extends beyond parks to the production of knowledge itself. The Agence Science-Presse describes a contradictory dynamic for 2026: AI can have a positive impact on research, but scientific work also has to learn to resist an avalanche of low-quality content—described as “mush,” according to the outlet.
That matters for conservation because decisions rely on studies, inventories and models. If the research pipeline gets saturated with fragile results, the promise of acceleration can backfire. Science-Presse points to “virtual scientists” capable of rapidly testing thousands of configurations, and says generative AI systems have been tested since 2024 in laboratories around the world. It also reports that in 2026, data is expected to emerge on the harmful impact of a proliferation of very low-quality work on scientific labor, according to Agence Science-Presse.
For conservation, the risk cuts both ways. Faster models can help explore scenarios and target field campaigns. But unverified results can steer scarce resources toward the wrong priorities. Quality control becomes operational: field validation, transparent methods, and the ability to distinguish a robust warning from an artificial signal.
Underlying it all is trust. Science-Presse describes a consensus that AI is “here to stay.” In that context, conservation has to build guardrails—not to slow innovation, but to keep speed from replacing rigor. Technology can reveal hidden extinctions, but it can also manufacture noise. Both forces are in play.
Species discovery and ecosystem tracking: big promise, strict conditions
AI is also being promoted as a way to accelerate species discovery and make ecosystem monitoring easier. Analyses on AI and nature protection point to its ability to speed steps from identification to classification and contribute to conservation goals, according to “L’intelligence artificielle, outil sous-exploité pour la protection de la …”.
The promise aligns with Kew’s thesis: fill in the shadows. If automated tools can more quickly spot signals of rarity, poorly known occurrences or localized declines, they can steer human effort toward what matters most. But the condition remains the same: baseline reference data, clear protocols, and the ability to explain results. Without that, AI becomes an error amplifier.
Field practitioners already know the test: sensors, images and algorithms only matter if they lead to workable decisions—adjusting site management, protecting breeding areas, limiting disturbance, or better documenting a species’ range. Kew places that move from detection to action at the center of its message, presenting technology as a way to identify where urgency is highest, according to Kew.
What emerges is a conservation world that is more data-driven, but not machine-run. Digital tools expand what’s visible; they don’t replace inventories, botanists, mycologists or site managers. The shift described in 2026 looks less like automation than a reorganization of effort: less time searching blindly, more time verifying, protecting and restoring where signals converge.
Sources
- L'IA au cœur de la conservation : entre illusion et protection
- La technologie au service de la nature : comment les parcs d’Espagne utilisent l’IA pour protéger les rapaces, les chauves-souris et les zones humides de montagne – Article | IUCN
- L'intelligence artificielle, outil sous-exploité pour la protection de la …
- 5 choses à surveiller sur l’IA en 2026 | Agence Science-Presse
- L'IA au secours de la biodiversité ? – Bamboo Rose
