Our Short-Lived Anthropocene and the Coming Algorithmocene

The Anthroposcream

The unthinkable is starting to look inevitable

Almost lost in the cacophony of terrible news for us humans this year have been the recent signals of irreversible damage we humans have been doing to the planet.

The melting of Greenland’s ice sheet has essentially passed the point of no return according to a study published this August in Nature. And as one of the study’s co-authors puts it “… there is obviously more to come”. In particular, there are now strong indications that other irreversible climate tipping points, like deforestation, permafrost thawing and slowing of ocean circulations, are rapidly being reached. All in all, more climate scientists are now seeing the International Panel on Climate Change’s worst-case “God help us” scenario, once almost universally dismissed as alarmist “climate porn”, as the base-case for the future[1].

The blood-orange skies and destroyed homes from historic wildfires in the US, like the recent ones in the Amazon and Australia, offer a grim preview of such a future. It’s also a grim future for wildlife and their habitat. The damage in the US this year is still ongoing, but in last year’s Australian fires alone, over a billion animals and just as many trees are believed to have perished. The Amazon’s ecosystem never evolved to cope with fire, but that is also now part of its future. As a result, the world’s largest rainforest and richest store of biodiversity is in danger of becoming a dry savannah within the next several decades. Overall, the planet has lost almost 70% of its wildlife since 1970, and possibly half of all species will be gone by the end of this century in what is now recognized as Earth’s 6th Great Extinction event. Here too, the once unthinkable is looking more like the inevitable.

Yet as surreal as it is to say it, climate change and species extinction are only a part of the seismic disruptions the Anthropocene — the epoch of humans — is creating this century. To my mind, they are probably the middle half. In order of what we can think of as their future “geological footprint”, the four biggest trends shaping the century are

• Unsustainable human demographics
• Climate change
• Species extinction
• Unrestricted artificial intelligence

Each of these trends alone is big and complex enough to count as its own “challenge of the century” for global civilization. So it’s understandable that they’re often discussed as if happening in separate silos. But of course they are anything but separate. Instead they’re deeply linked in a mutual feedback loop entangling the entire triad of society, the environment and technology. As such, their convergence to a perfect storm this century is not a coincidence. They are also driven by socioeconomic rules as old as the Neolithic revolution. That, to my mind at least, essentially guarantees their continuation no matter how many warning lights are flashed in our faces.

As humankind blithely stares down the barrel of the truly unthinkable, one possible outcome this century is the infamous Great Filter: a collapse of global civilization and probably human populations. For all we know, this may even be the most likely outcome. But like many, I wouldn’t equate this scenario with the outright extinction of our species or even our technology. The minimum viable human population needed to avoid extinction is estimated to be less than 1,000 individuals. That could readily be managed, for example, in a few self-sustaining cave systems. Especially if, as expected under such extreme conditions, a disproportionate number of the survivors are engineers.

But if the survivors do eventually repopulate the planet, or if civilization manages to avoid collapse in the first place, one can argue it can only be with a very different dynamic between human society, the environment and technology. So different that, however this century plays out, the term Anthropocene as a new epoch in Earth history may not be as fitting as another: the Algorithmocene.

The 21st century in the context of Earth history

Before going there, let’s understand the magnitude of each of the above four disruptions in the context of their future “geological footprint” on Earth.

Like a symphony is organized into movements, sections, phrases and measures, Earth’s 4.5 billion year history is organized by geologists into nested time intervals of eons, eras, periods and epochs (ignoring for simplicity the shortest interval of ages). Specifically, eons (10⁹-10⁸ years) are divided into eras(10⁸-10⁷ years) which are divided into periods (10⁸-10⁶ years) which are divided into epochs (10⁷-10⁴ years).

From www.stratigraphy.org

What we think of as human civilization began with the Neolithic revolution and has happened wholly within the

• Holocene epoch — the last 11,000 years since the end of the last ice age
• Quaternary period — the last 2.5 million years in which climate swings have been driven by the Milankovitch cycles of Earth’s orbit and rotation
• Cenozoic era — the last 65 million years since a stray asteroid ended the reign of the dinosaurs and paved the way for the age of mammals
• Phanerozoic eon — the last 540 million years since the Cambrian explosion created the fantastic variety of complex life we have had so far.

One peculiar thing of note is that each of the above is the shortest in its respective time unit. That is, our Holocene is the shortest of all of Earth’s 38 epochs. Similarly, our Quaternary, Cenozoic and Phanerozoic are the shortest among Earth’s 22 periods , 10 eras and 4 eons respectively. If we imagined selecting these as independent random samples[2], the odds of getting this streak would be about 0.003%; the same odds as getting a royal flush in 7-card Texas hold ’em. No doubt this just an odd coincidence (somewhat like “Bode’s Law” in astronomy shows odd coincidences in the solar system’s planetary orbits). But if we pretend for a moment that it isn’t, the trend would predict something interesting (just as Bode’s Law predicted Neptune’s orbit). Namely, that our current Holocene epoch should be ending sooner than later.

And indeed, we know the Holocene is ending even faster than sooner. Sometime in the next year or so, the International Commission on Stratigraphy is due to officially introduce the long awaited Anthropocene — the epoch of humans — into Earth’s history book [3].

In keeping with the above imagined trend, the changes our species has set in motion go much further than moving to the next epoch. Specifically, we can see the following links

• Human demographics is ending the Holocene. That’s of course already in the new name Anthropocene — the epoch of humans. But it’s not our sheer numbers per se but the way we live that has changed everything. Namely,
• Climate change is ending 2.5 million years of the Quaternary period. Due to the now disrupted Milankovitch cycles of climate swings, for example, it’s estimated that the next ice age has been delayed by at least 100,000 years. But even this disruption is not as long lived as the next one.
• Species extinction is ending 65 million years of the Cenozoic era. Great extinction events have marked the boundaries of geologic eras since the Cambrian. Our Cenozoic, meaning “new life” began with the vanishing of 75% of all species in Earth’s 5th Great Extinction event. In the future fossil record, it should be bracketed by our current 6th Great Extinction.

Disconnected from the reality of our situation

Before going further down the “geological footprint” list, let’s pause for a moment and notice something important about ourselves. The fact that even our “normal” civilized activity (i.e. sans nuclear war or similar) is upending the last 65 million years of Earth’s history should leave us all stupefied. Such a terrifying tally should lead us, as rational, civilized beings, to an immediate and complete overhaul of our daily life and the socioeconomic rules that drove us here in the first place.

But of course nothing of the kind is happening, nor can we expect it to. After all, we still build homes in the middle of fire hazard zones and floodplains. Even during this pandemic, otherwise rational people are happy to gather in large, maskless crowds. More than 50 years after the first official warnings at the highest levels of government and industry, global carbon emissions still track one thing: global GDP. In short, the apocalypse remains abstract and remote for as long we think we’re able to personally avoid its consequences. As climatologist Ken Caldeira has remarked in the Guardian: “If people are rich enough to air-condition their lives, they can watch whatever is the successor to Game of Thrones on TV, as the natural world decays around them”.

Case in point: even after making this tally, I haven’t changed my daily routine. Sure, for a Western urbanite I try to behave in (what passes here for) an “environmentally conscious” way. But only until that hits an actual pain point. Things like recycling, public transport and buying local-organic are fortunately easy where we live. But I still binge on my favorite TV series and Youtube channels with a carbon footprint of 10 kg per hour; and I certainly won’t give up my five cups of coffee a day, each with a water footprint of 130 liters per cup.

That’s not because I don’t believe in the dangers I’m writing about, but because that’s just how we humans are. At least that’s how humans in technological societies are. From the “bottom of the pyramid” to the “top 1%” and across the political belief spectrum, the vast majority of us will not leave our behavioral comfort zones until we’re blasted out with dynamite. Yes, there are deep-seated socio-psychological reasons for this. But it’s also because technological civilization is based on a simple premise. Even when faced with problems of our own making, why improve our behavior when we can improve our technology instead?

The Algorithmocene is coming

Which brings us to the existential irony of this point in human history. The problems created by our technological civilization are so massive and approaching so fast that only a virtual miracle can now avert the worst. If it can be averted at all. Blaming callous 1%-ers and their political cronies, socioeconomic inequality, poor education, wasteful consumer culture, human moral failing in general and even the Neolithic revolution itself may all have their proper place. But these debates are also centuries old and not going to move any needles anytime soon. Wishful thinking aside, at this point in the 21st century we are simply out of time to improve human behavior. The only miracle left to hope for is technology.

To be clear, that’s not intended as a cornucopian argument as much as a fatalistic one. The risk of this miracle cure being worse than the disease is real. And it may be too late anyway: even if “innovation is the infinite resource”, as Ramez Naam claims, that’s irrelevant when time is finite. Every collapsed civilization in history is a testament to that fact.

Either way, having broken Earth’s natural equilibria, our global 21st century civilization will try its technological best to impose its artificial equilibria. And as technologies go, artificial intelligence has the greatest potential of all to fall in that category.

AI is already driving the greatest technological arms race in history. Businesses and nations great and small already see AI as the ultimate strategic investment in their own future. From finance, design, manufacturing, distribution, and transportation to agriculture, politics, law-enforcement, work-life, social life, health and entertainment, AI is already transforming how our world works. It’s also (one might say unsurprisingly) transforming how our very hackable minds work.

Yet it’s also clear to everyone that the story of artificial intelligence on planet Earth has barely begun. Its capabilities grow at least as fast as Moore’s Law, and the transformation of every corner of life is already baked in. Whatever fears (founded or unfounded) of a Terminator world we may have: the scale and complexity of navigating this century will, to my mind, ensure that AI will become the dominant factor shaping the world going forward. So much so that the concept of AI as a “human tool”, I argue, will eventually lose its meaning.

But before that sets off any woo-woo alarms: none of the above is a claim about the “singularity”, AI “consciousness” or even general artificial intelligence (all of which I’m skeptical for different reasons). It’s rather a claim about where the information processing action on planet Earth will be. And about how that action should be understood.

As the overwhelming mass of the world’s high-level information flow shifts from humans to machines, and algorithms get better at hacking into our mental space, an obvious question emerges. Where in the array of human affairs does any causal chain of events begin? How meaningful is the very concept of “human intention” in such a world? One might pretend that at least human programmers and their intentions are still behind it all, but in the age when algorithms pre-process everyone’s information and options, is this not merely a distinction without a difference?

In short, if visiting aliens survey our world next century and are able to “follow the money” — i.e. the flow of information — they may well decide they’re looking at a world run by machine-based algorithms. Whether they will pay much attention to the two-legged carbon-based creatures servicing them is another question. After all, any visiting aliens may well be machines from their own Algorithmocene age.

It’s worth noting that for any future geologists, our Algorithmocene should leave a clear signal in the fossil record. Already there are more IoT devices in the world than humans, and their numbers are set to skyrocket in the coming decades. Given the 50 million tons of electronic waste the world produces every year, the fossil record may even already have it’s so-called golden spike: Agbogbloshie on the outskirts of Accra, Ghana. Agbogbloshie is said to be home to the world’s largest e-waste dump, nicknamed Sodom for its constant dumpster fires. As filmmaker Florian Weigsamer described it in a 2018 interview with CBC Radio “When you first come there, it looks like an apocalyptic place…it seems like a place where everything ends…”

Image from the Blacksmith Institute for a Pure Earth, a non-profit working to improve recycling conditions in Agbogbloshie and elsewhere.

Epilogue: what the Algorithmocene might look like

For the next few decades at least, it’s hard to see how the future will not continue to devolve into ever more terrible versions of the present. Everything trending badly in the world now is doing so for reasons much bigger than any conceivable public policy tweaks that might mitigate them. But for the longer term towards the end of the century, three possible scenarios seem conceivable enough to describe briefly in terms of famous sci-fi memes.

• The utopia-scenario means we somehow avoid catastrophe with the planet’s climate and ecosystems. That means engineering on a planetary scale. Unfortunately, our current competition-based system of nation-states and capitalism has no political or business model for managing a whole planet. Hence a world-government system like in Star-Trek would probably be needed. (Here it may be worth remembering that Star Trek never explicitly mentions democracy, but so be it.) Either way, managing such astronomical complexity means operational control of utopia will be via autonomous AI systems. What will happen to human minds and societies in such a scenario can only be guessed at. Perhaps a true utopia of peace, prosperity and the fulfillment of human potential could be reached. But then again, this scenario is no doubt the least realistic one.
• The middle scenario is an AI-powered version of Mad Max: an inhospitable, conflict-ridden planet of anarchy and technology. The main driver of this scenario is straight-forward and to my mind determinative: the certainty that autonomous weaponized AI systems are coming. And they will be ubiquitous. The world has been lucky that standard WMDs (nuclear, chemical and biological) can be limited through the materials and massive infrastructure they require. Treaties here are at least somewhat enforceable. Future AI weapons can be assembled with off-the-shelf components in the proverbial garage laboratory. And no military or para-military group in the world can afford to be on the wrong side of this development curve. Because this scenario only requires a small number of humans to retain the same behavior we’ve had for all of history, this is no doubt by far the most realistic one.
• The final horror scenario I personally worry about more than a murderous, self-aware Terminator is another meme from the early 1990s: the “grey goo” apocalypse. Nanobots with AI-powered swarm intelligence eating their way through the world sounds almost quaint in 2020. And to be sure, it’s technically still some distance away. On the one hand, swarm intelligence is much cheaper and easier to implement than general intelligence (which seems to me to be even further in the future). After all, nature made the human brain only once, but has been doing swarm intelligence for eons. On the other hand, nanoscale components like actuators and controllers are still quite difficult. Perhaps we should all hope they stay difficult. Otherwise, our species may yet succeed in erasing the Phanerozoic eon of the last 540 million years. Technically then, the proper nomenclature for that eon would be the Algorithmozoic.

Footnotes:

[1] To dispel any illusions about the world effectively addressing climate change, see NOAA’s atmospheric carbon dioxide monitoring site. Even the latest UN report on renewables 2020 states that current global commitments to reduce carbon “…fall far short of what would be needed to limit world temperature increases to less than 2 degrees Celsius.”

[2] They are of course neither independent nor random. But even if the estimate is off by an order of magnitude, it’s an impressive coincidence.

[3] The term “Anthropocene” was widely popularized at the beginning of this century by Paul Crutzen and others since. However the International Committee on Stratigraphy is an extremely deliberative body. It is set to formally review the adoption of the term in 2021.