Mass extinctions are meant to arrive with a bang. Rock from space, sky on fire, one appalling afternoon written permanently into the sediment. The event many biologists say is underway right now looks nothing like that. It shows up as chainsaws, trawl nets, shipping lanes and a slow chemical drift in the air. It has been running long enough that nobody alive remembers what the baseline looked like.

Where the bar actually sits

Palaeontologists have a threshold, and it is a brutal one. Losing more than three-quarters of species in a geologically short window is what qualifies, and by that measure, it has happened five times in roughly the past 540 million years. Anthony Barnosky and colleagues at UC Berkeley ran the modern numbers against the fossil record in a 2011 review in Nature and found current extinction rates sit well above the ordinary background hum, while the total body count so far falls short of the Big Five. Their projection was blunt. If every species currently classed as threatened disappeared, and that pace held, a sixth mass extinction could arrive within three to twenty-two centuries.

By the standards of the fossil record, a few centuries is a blink.

Nobody has counted the invertebrates

The IUCN Red List, extinction’s standard scorecard, has assessed nearly every bird and mammal on the planet and only a sliver of the invertebrates, which is where almost all biodiversity actually lives. That imbalance is the argument’s soft spot. Robert Cowie and co-authors at the University of Hawai’i argued in the journal Biological Reviews that the bias quietly deflates the official figures. Extrapolating from land snails and slugs, a group taxonomists know reasonably well, they estimated that between 7.5 and 13 per cent of roughly two million known species may already be gone since about 1500. That works out to somewhere between 150,000 and 260,000 species. The Red List’s formal tally of confirmed extinctions was 882.

That’s an enormous gap, and it rests on an extrapolation the authors themselves called bold. One paper, one method, one phylum of snails standing in for the rest of the invertebrate world.

Speed is the part that stands out

Rate is what separates a mass extinction from ordinary churn, since species have always died out at a low, steady drip. Gerardo Ceballos’s team stacked the assumptions against their own conclusion in a paper published in Science Advances, adopting a background rate twice as high as the standard estimate. Vertebrates came out vanishing up to 100 times faster than that conservative baseline would predict. Under that baseline, the species lost over the last century would otherwise have taken between 800 and 10,000 years to disappear.

The chemistry is moving at a similar clip. Richard Zeebe, working with Andy Ridgwell and James Zachos, reconstructed the carbon release behind the Palaeocene-Eocene Thermal Maximum 56 million years ago, the fastest natural carbon spike of the last 66 million years, and published the result in Nature Geoscience. That ancient surge, violent enough to acidify the oceans and kill off deep sea foraminifera, ran at less than a tenth of the current human rate. Their term for where this leaves the present day is a “no-analogue” state, meaning there is nothing in that stretch of geological history to check the forecast against.

What is actually doing the killing

No asteroid, no supervolcano, no obvious villain. When IPBES ranked the direct drivers of change in nature by global impact, the top two were changes in how land and sea are used, followed by hunting, fishing and logging. Climate change came third, which surprises people who assume warming is the whole story. The Living Planet Report points the same way, naming habitat loss and degradation as the most reported threat to wildlife populations worldwide, driven primarily by how the world grows its food, with overexploitation, invasive species and disease behind it. Freshwater populations have fared worst of all, down 85 per cent on the same measure.

The mechanisms are dull, but the fallout is measured in geological time.

What the big numbers do and do not say

Ask around and two figures come back. One million species threatened with extinction, from the IPBES global assessment in 2019. A 73 per cent decline in wildlife populations since 1970, from the WWF and Zoological Society of London Living Planet Index.

Both are real, and both get mangled in transit. The 73 per cent describes an average change in the size of monitored populations, not a headcount of animals lost, and Our World in Data has walked through the arithmetic showing how a small number of catastrophic crashes drag that average far below the experience of the typical population. IPBES, for its part, pointedly avoided the term mass extinction. In testimony to the US Congress summarising the assessment, IPBES chair Robert Watson noted that losing all one million threatened species would still leave the tally nowhere near the 75 per cent mark.

Working scientists are still scrapping over it. Cowie returned to the argument in a 2025 piece in Trends in Ecology & Evolution, conceding the threshold has not been crossed while arguing that framing the risk as speculation makes it easier for governments to do nothing about it.

An extinction with witnesses

Every previous mass extinction ran its course unobserved. Trilobites did not convene panels. This one has an audience, a reasonable set of instruments, and a live argument about definitions playing out while the numbers keep moving underneath it.

Whether any of that changes the outcome is still an open question. The five extinctions before this one never got a vote. This one might.