The Wild Within Our World: How Technology Is Revealing Nature We Still Don’t Know

---

Much of the natural world is still missing from our maps, databases and scientific records not because it has been deliberately hidden, but because humanity has never finished documenting it.

That gap matters more than it may seem. Scientists cannot protect a species they do not know exists, assess an extinction they cannot detect, or understand how climate change is reshaping an ecosystem without enough evidence about what lives there. In 2026, however, a new set of tools is beginning to change the scale of that problem.

The Royal Botanic Gardens, Kew’s State of the World’s Plants and Fungi 2026 report describes a shift in biodiversity science driven by digitisation, artificial intelligence, DNA sequencing and global data-sharing. More than 400 scientists from over 170 institutions in 40 countries contributed to the report. Its central message is both encouraging and unsettling: technology can help us see more of the living world, but it is also revealing how much we still do not know.

Key Takeaways

• Scientists have assessed only a fraction of known plant and fungal species for extinction risk.
• More than 100,000 plant species and over 2 million fungal species are estimated to remain unknown to science.
• Digitised specimens allow researchers to study biodiversity at a scale that was previously difficult or impossible.
• AI can accelerate species identification and data analysis, but expert verification remains essential.
• Historical specimens are becoming climate records, revealing changes in plant flowering patterns over the past century.
• The biggest challenge is no longer simply collecting information, but making biodiversity data more complete, representative and accessible.

We Know Less About Nature Than We Think

The familiar image of biodiversity often begins with a catalogue: species names, locations, population estimates and conservation status.

But the catalogue is incomplete.

Kew’s 2026 assessment reports that 29,748 plant species and 411 fungal species are currently listed as threatened with extinction. Yet only about 18% of known plant species and less than 1% of fungi have been assessed for extinction risk. That means the numbers we use to describe the biodiversity crisis are themselves constrained by what scientists have had the opportunity to study.

The problem becomes larger when unknown species are considered.

Kew estimates that more than 100,000 plant species and more than 2 million fungal species remain unknown to science. Thousands of additional plant and fungal species were formally named in 2024 and 2025, but identifying species remains a race against time because some may disappear before scientists have fully documented them.

This creates an uncomfortable paradox: humanity is discovering new forms of life while some forms of life may already be disappearing.

The Museum Specimen Is Becoming a Digital Sensor

For centuries, botanists and mycologists have preserved specimens in herbaria and fungaria. These collections can look like historical archives pressed plants, dried fungi, handwritten labels and carefully recorded locations.

Digitisation changes their role.

A specimen collected decades or even centuries ago can contain information about where a species lived, when it was flowering, what its physical characteristics were and how scientists understood it at the time. Once photographed, catalogued and connected to other records, thousands or millions of specimens can be analysed together.

Kew’s 2026 report notes that the world’s herbaria contain roughly 406 million specimens, yet fewer than 16% have been imaged and made available online. The digital record of life is therefore still dramatically smaller than the physical record stored in scientific collections.

That distinction is important.

A specimen locked inside a collection can be studied by someone who can physically reach it. A digitised specimen can potentially become part of international research, provided the associated data are accessible and reliable.

The result is a transformation in scale: instead of examining one specimen at a time, researchers can begin asking questions across entire collections.

AI Is Helping Scientists Search the Biological Haystack

Artificial intelligence does not make biodiversity research automatic. Its value is more practical.

Species identification can involve subtle differences that require specialist knowledge. Digitised specimen collections also contain enormous amounts of information that would take humans years to process manually.

Machine-learning systems can help classify images, identify patterns and prioritise specimens for closer examination. Researchers are also exploring large language models to transcribe information from specimen labels, converting historical records into machine-readable data.

This could allow scientists to spend less time performing repetitive tasks and more time investigating unusual or potentially important cases.

But there is a critical limitation.

AI systems learn from the information provided to them. If the underlying collections are geographically biased, poorly labelled or incomplete, an algorithm can reproduce those weaknesses at much greater speed.

Kew therefore stresses that AI should assist rather than replace scientific expertise. Human researchers still need to validate results, interpret uncertain cases and determine whether an apparent pattern reflects biology or simply a gap in the data.

The lesson extends well beyond conservation: faster analysis is useful only when the underlying evidence is good enough to analyze.

Old Specimens Are Revealing a Changing Climate

One of the most intriguing possibilities of digitised biodiversity collections is that they can turn historical specimens into records of environmental change.

Kew reports that researchers used AI to analyse approximately eight million plant specimens and found that flowering times have shifted by an average of 2.5 days per decade over the past century. The changes are not uniform across the planet, and tropical regions showed particularly important patterns involving both temperature and rainfall.

That matters because flowering is connected to much more than the appearance of a plant.

Plants interact with pollinators, herbivores and other organisms through seasonal cycles. When those cycles shift, ecological relationships can also change.

The significance of digitisation here is subtle. The technology did not create the historical evidence. The evidence was already sitting inside collections.

Digital tools made it possible to connect enormous numbers of old observations and examine them as a global dataset.

In that sense, technology is not replacing nature research. It is making previously inaccessible evidence usable.

The Extinction Problem Has a Blind Spot

There is another difficulty that technology may help address: proving that something has disappeared.

Declaring a species extinct sounds straightforward, but absence is difficult to demonstrate. A plant may not have been observed for decades because its habitat is remote, surveys are inadequate, or the species is extremely rare.

Kew’s 2026 report highlights what it calls the Katuš shortfall the gap created when biodiversity disappears without the loss being properly recorded. Scientists are increasingly exploring probability models that combine historical records, digitised specimens and sighting data to estimate whether a species is genuinely extinct or simply undetected.

This approach could change conservation priorities.

Instead of treating extinction as a simple binary question alive or extinct researchers can estimate degrees of uncertainty and identify species whose disappearance is becoming increasingly probable.

That is particularly important for organisms that receive less scientific attention than charismatic animals.

The Fungal World May Be the Biggest Unknown

Plants are relatively visible. Fungi are not.

Kew’s earlier research estimated that there may be around 2.5 million fungal species globally, while only about 155,000 had been formally named at the time of the 2023 assessment. More than 90% of fungal diversity therefore remained undescribed.

The 2026 report describes another technological opportunity: scientists are extracting and analysing DNA from fungal specimens that have been preserved for decades and, in some cases, nearly two centuries.

That opens a biological archive that previously could not be examined in the same way.

Fungi matter because they are deeply connected to ecosystems, agriculture and human health. Their genetic diversity may also contain organisms or biological mechanisms with potential applications in medicine, crop protection and other fields.

The important point is not that every undiscovered fungus will produce a useful invention. There is no evidence for such a guarantee.

The point is that a vast biological resource remains poorly understood and technology is beginning to make that resource scientifically accessible.

Technology Cannot Solve a Data Problem by Itself

The promise of digital biodiversity science comes with an important warning.

The world’s biological collections are not evenly distributed or digitised. Kew reports that digitised herbaria remain concentrated in the Global North, leaving major gaps in knowledge about biodiversity in other regions, including areas that may contain exceptional biological diversity.

That creates a potential feedback loop.

If researchers have more digital records from some countries than others, models trained on those records may produce a distorted picture of global biodiversity. An algorithm can process billions of observations and still produce an incomplete answer if the underlying geography is incomplete.

Digitisation therefore needs to be accompanied by investment in collections, local scientific capacity, standardised data practices and equitable access to biodiversity information.

The technology is powerful. The data ecosystem surrounding it matters just as much.

A New Way of Seeing the Wild

The most important development may not be any individual AI model, database or scanning project.

It is the changing relationship between physical evidence and digital knowledge.

A pressed leaf collected generations ago can now become part of a global dataset. A fungal specimen stored for decades can yield genetic information. A handwritten label can be converted into searchable data. Millions of observations can be compared computationally. Mathematical models can help estimate whether species have disappeared without leaving a clear record.

These developments do not eliminate the biodiversity crisis.

They improve our ability to see it.

That distinction matters because conservation begins with knowledge. Before a species can be protected, someone needs to know that it exists, where it occurs, what threatens it and what evidence shows that its population is changing.

The wild within our world is therefore larger than the catalogue suggests.

Technology is giving scientists a better chance of discovering what has been overlooked but the discoveries also make the responsibility clearer. Knowing more about nature is valuable only if that knowledge is translated into better decisions about what humanity chooses to protect.

Conclusion

The next frontier in biodiversity conservation may not be somewhere unexplored in a distant rainforest. Some of it is already sitting in drawers, cabinets, laboratories and collections around the world.

The challenge is to make those records visible, connect them responsibly and use them to understand a living planet that remains only partially documented.

Kew’s 2026 findings offer a cautiously hopeful message. AI, digitisation, DNA sequencing and data-sharing can accelerate discovery and reveal ecological changes that were previously difficult to see. But they cannot compensate for missing collections, biased datasets or insufficient scientific capacity.

The future of conservation may therefore depend on a deceptively simple principle: we cannot protect the wild we have not yet learned to see.