Calum Chace's Blog, page 15

The Simulation Hypothesis: an economical twist (part 1 of 2)

Are we living in the Matrix?

Are we living in the Matrix? This question seems futuristic, a theme from a science fiction movie. (Which of course it is.)  But the best science fiction is philosophy in fancy dress, and philosophers have been asking the question since at least the ancient Greeks.

The question generated considerable interest in June this year when Elon Musk said the chance that we live in a “base reality” was “one in billions”. But as long ago as 380 BC, when the Greek philosopher Plato wrote “The Republic”, he included the Allegory of the Cave, which argued that men are like prisoners chained to a wall in a dark cave. Their only impression of reality is the shadows cast on a wall in front of them by a fire behind them which they cannot see. If they could see reality in all its true glory it would blind them.

Trilemma

In 2003, the question was given the formal structure of a trilemma (a difficult choice between three options) by the Oxford University philosopher Nick Bostrom. He observed that we use our technology to simulate worlds inhabited by creatures which are as realistic as we can make them, and that some of the most compelling of these worlds contain people who could be our ancestors. There are many reasons to create these “ancestor simulations”, and a civilisation which could create them at all would create many of them. He concluded that one of these three statements must be true:

All civilisations are destroyed (or self-destruct) before their technology reaches the stage where it would allow them to create ancestor simulations.
Numerous civilisations have reached that stage of technology but they all refrain from doing so, perhaps for moral or aesthetic reasons.
We live in a simulation. (1)

People often that assume that Bostrom argues that we are living in a simulation, but in fact he thinks the self-destruction scenario is also a contender.

Answering Enrico

The simulation scenario and the inevitable demise scenario also both offer plausible answers to Enrico Fermi’s Paradox, the apparent contradiction between the absence of evidence of other intelligent life in the universe, and the high probability of there being lots of it. Maybe civilisations inevitably collapse before they acquire the power to build Dyson Spheres or blurt out Pi in a durable pan-galactic email. Or maybe, like Douglas Adams’ mice (2), the creators of the simulation we live in only needed the one world to achieve whatever goal they had in creating the earth. Populating the universe with teeming swarms of aliens was simply unnecessary, as well as being hideously expensive in computational resources.

Reasons to simulate us

If we do live in a simulation, why was it made? If the sole purpose was entertainment then the prevalence in our world of torture, child leukemia and much else makes it hard to avoid concluding that our creator was either incompetent or callous. (Conscientious theists have to wrestle with a fiercer version of this problem known as the Problem of Evil. I say fiercer because in their case the incompetent or callous entity is also omniscient, omnipresent and omnipotent, which makes their world a very scary place. Especially since most theists believe in a deity which actively intervenes in their world from time to time.)

Perhaps we are part of a massive modelling exercise (a Monte Carlo simulation, perhaps) to determine the optimal way of achieving a specific goal. One candidate for that goal is the creation of superintelligence. From the perspective of a superintelligent alien species, individual humans might be about as interesting as dogs if we are lucky, or as microbes if not. Whereas a superintelligence might be seriously interesting. Maybe our simulators are in the business of creating new friends. Or maybe they need a lot of clever new colleagues to solve a really big problem, like the heat death of the universe.

Next week: the economical twist

Notes

(1) Bostrom’s trilemma makes a few assumptions, notably:

It is not impossible to create simulations inhabited by entities which consider themselves to be alive.
It is implausible that we are the most (or one of the most) technologically advanced civilisations that have ever existed in this universe.
It is implausible that we are one of the tiny, tiny minority of civilisations in this universe which arose naturally.

(2) The Hitchhikers Guide to the Galaxy

The Hawking Recursion*

A couple of years ago Stephen Hawking told us that (general) AI is coming, and it will either be the best or the worst thing to happen to humanity.  His comments owed a lot to Nick Bostrom’s seminal book “Superintelligence” and also to Professor Stuart Russell.  They kicked off a series of remarks by the Three Wise Men (Hawking, Musk and Gates), which collectively did so much to alert thinking people everywhere to the dramatic progress of AI.  They were important comments, and IMHO they were r

Journalists are busy people, good news is no news, and if it bleeds it leads.  For a year or so the papers were full of pictures of the Terminator, and we had an epidemic of Droid Dread: the Great Robot Freak-Out of 2015.

The Terminator is an arresting image, and it’s good for getting people’s attention, but making people unreasonably scared of AI is unhelpful in the medium term.  Among other things, it makes sensible AI resarchers cringe, and in turn they may under-state the risks inherent in AI.  Or, like senior Googlers, it makes them go silent on the subject.

So when Professor Hawking repeated his remarks during the launch of the Leverhulme Centre for the Future of Intelligence in Cambridge this week (October 2016), there was a need for balanced comment.  Here is my contribution, in an interview with Sky News.

*If you’re wondering whether this title owes anything to The Big Bang Theory, you’re right.

 

Book review: “Homo Deus” by Yuval Harari, (part 3 of 3)

Week Three: Extreme Algocracy

In the first third of Homo Deus, Harari claims that humanity has more-or-less conquered its traditional enemies of famine, plague and war, and has moved on to chasing immortality, happiness and divinity. He also urged us all to become vegetarian, if only to make us less contemptible in the eyes of a future superintelligence.

In the second third of the book he offers a surprisingly cursory review of the two singularities – the economic and the technological. He seems to assume that most people already know about them, or at least won’t need much persuading to accept them.

Extreme algocracy

So finally we arrive at the main event, in which Harari predicts the dissolution of not only humanism, but of the whole notion of individual human beings. “The new technologies of the twenty-first century may thus reverse the humanist revolution, stripping humans of their authority, and empowering non-human algorithms instead … Once Google, Facebook and other algorithms become all-knowing oracles, they may well evolve into agents and finally into sovereigns.” As a consequence, “humans will no longer be autonomous entities directed by the stories their narrating self invents. Instead, they will be integral parts of a huge global network.” This seems to be an extreme version of an idea called algocracy, in which humans are governed by algorithms.  The philosopher John Danaher is doing interesting work on this.

As an example of how extreme algocracy could come about, Harari asks you to “suppose my narrating self makes a New Year resolution to start a diet and go to the gym every day. A week later, when it is time to go to the gym, the experiencing self asks Cortana to turn on the TV and order pizza. What should Cortana do?” Harari thinks Cortana (or Siri, or whatever they are called then) will know us better than we do and will make wiser choices than we would in almost all circumstances. We will have no sensible option other than to hand over almost all decision-making to them.

Two new religions

Given his religious turn of mind, it is no surprise that Harari sees this extreme algocracy as leading to the birth of not one, but two new religions. “The most interesting place in the world from a religious perspective is not the Islamic State or the Bible Belt, but Silicon Valley.” Algocracy, he thinks, will generate two new “techno-religions … techno-humanism and data religion”, or “Dataism”.

“Techno-humanism agrees that Homo Sapiens as we know it has run its historical course and will no longer be relevant in the future, but concludes that we should therefore use technology in order to create Homo Deus – a much superior human model.” Harari thinks that techno-humanism is incoherent because if you can always improve yourself then you are no longer an independent agent: “Once people could design and redesign their will, we could no longer see it as the ultimate source of all meaning and authority. For no matter what our will says, we can always make it say something else.” “What”, he asks, “will happen to society, politics and daily life when non-conscious but highly intelligent algorithms know us better than we know ourselves?”

Dataism trumps Techno-humanism

“Hence a bolder techno-religion seeks to sever the humanist umbilical cord altogether. … The most interesting emerging religion is Dataism, which venerates neither gods nor man – it worships data. … According to Dataism, King Lear and the flu virus are just two patterns of data flow that can be analysed using the same basic concepts and tools. … Humans are merely tools for creating the Internet-of-All-Things, which may eventually spread out from planet Earth to cover the whole galaxy and even the whole universe. This cosmic data-processing system would be like God. It will be everywhere and will control everything, and humans are destined to merge into it.”

“Dataism isn’t limited to idle prophecies. Like every religion, it has its practical commandments. First and foremost, a Dataist ought to maximise data flow by connecting to more and more media, and producing and consuming more and more information.” So people who record every aspect of their lives on Facebook and Twitter are not clogging up the airwaves with junk after all; they are simply budding Dataists.

I am un-persuaded by the idea that mere data, undifferentiated by anything beyond quantity and complexity could become sovereign on this planet and throughout the universe. I think Harari has missed an interesting opportunity – if he replaced the notion of data with the notion of consciousness, I think he might be onto something important. It would not be the first time that a thinker proposed that mankind’s destiny (a religiously loaded word which he would perhaps approve) is to merge its individual minds into a single consciousness and spread it throughout the cosmos, but it might be the first time that a genuinely mainstream book did so.

In any case, Harari deserves great credit for staring directly at the massive transformations heading our way, and following his intuitions to their logical conclusions.

TL;DR:

A rather plodding first half may deter some fans of “Sapiens” (Harari’s previous book), but it is worth persevering for the extreme views about algocracy which he introduces in the final third.

Book review: “Homo Deus” by Yuval Harari, (part 2 of 3)

Week two: the two Singularities

In the first third of his new book, Yuval Harari described how humanity escaped its perennial evils of famine, plague and war, and he claimed that we are now striving for immortality, happiness and divinity.  Now he enters the territory of the economic and the technological singularities.  Read on…

Free will is an illusion

Harari begins this section by attacking our strong intuitive belief that we are all unitary, self-directing persons, possessing free will. “To the best of our scientific understanding, determinism and randomness have divided the entire cake between them, leaving not even a crumb for ‘freedom’. … The notion that you have a single self … is just another liberal myth, debunked by the latest scientific research.” This dismissal of personal identity (the “narrating self”) as a convenient fiction will play an important role in the final third of the book, which we will get to next week.

A curious characteristic of “Homo Deus” is that Harari assumes there is no need to persuade his readers of the enormous impact that new technologies will have in the coming decades. Futurists like Ray Kurzweil, Nick Bostrom, Martin Ford and others (including me) spend considerable effort getting people to comprehend and take into account the astonishing power of exponential growth. Harari assumes everyone is already on board, which is surprising in such a mainstream book. I hope he is right, but I doubt it.

Stop worrying and learn to love the autonomous kill-bot

Harari is also quite happy to swim against the consensus when exploring the impact of these technologies. A lot of ink is currently being spilled in an attempt to halt the progress of autonomous weapons. Harari considers it a waste: “Suppose two drones fight each other in the air. One drone cannot fire a shot without first receiving the go-ahead from a human operator in some bunker. The other drone is fully autonomous. Which do you think will prevail? … Even if you care more about justice than victory, you should probably opt to replace your soldiers and pilots with autonomous robots and drones. Human soldiers murder, rape and pillage, and even when they try to behave themselves, they all too often kill civilians by mistake.”

The economic singularity and superfluous people

As well as dismissing attempts to forestall AI-enabled weaponry, Harari has no truck with the Reverse Luddite Fallacy, the idea that because automation has not caused lasting unemployment in the past it will not do so in the future. “Robots and computers … may soon outperform humans in most tasks. … Some economists predict that sooner or later, un-enhanced humans will be completely useless. … The most important question in twenty-first-century economics may well be what to do with all the superfluous people.”

My income is OK, but what am I for?

Harari has interesting things to say about some of the dangers of technological unemployment. He is sanguine about the ability of the post-jobs world to provide adequate incomes to the “superfluous people”, but like many other writers, he asks where we will find meaning in a post-jobs world. “The technological bonanza will probably make it feasible to feed and support the useless masses even without any effort on their side. But what will keep them occupied and content? People must do something, or they will go crazy. What will they do all day? One solution might be offered by drugs and computer games. … Yet such a development would deal a mortal blow to the liberal belief in the sacredness of human life and of human experiences.”

Personally, I think he has got this the wrong way round. Introducing Universal Basic Income (or some similar scheme to provide a good standard of living to the unemployable) will probably prove to be a significant challenge. Persuading the super-rich (whether they be humans or algorithms) to provide the rest of us with a comfortable income will, I hope, be possible, but it may have to be done globally and within a very short time-frame. If we do manage this transition smoothly, I suspect the great majority of people will quickly find worthwhile and enjoyable things to do with their new-found leisure. Rather like many pensioners do today, and aristocrats have done for centuries.

The Gods and the Useless

I have more common ground with Harari when he argues that inequality of wealth and income may become so severe that it leads to “speciation” – the division of the species into completely separate groups, whose vital interests may start to diverge. “As algorithms push humans out of the job market, wealth might become concentrated in the hands of the tiny elite that owns the all-powerful algorithms, creating unprecedented social inequality.”

Pursuing this idea, he coins the rather brutal phrase “the Gods and the Useless”. He points out that in the past, the products of technological advances have disseminated rapidly through economies and societies, but he thinks this may change. In the vital area of medical science, for instance, we are moving from an era when the goal was to bring as many people as possible up to a common standard of “normal health”, to a world in which the cognitive and physical performance of certain individuals may be raised to new and extraordinary heights. This could have dangerous consequences: when tribes of humans with different levels of capability have collided it has rarely gone well for the less powerful group.

The technological singularity

The technological singularity pops up briefly, and again Harari sees no need to expend much effort persuading his mainstream audience that this startling idea is plausible. “Some experts and thinkers, such as Nick Bostrom, warn that humankind is unlikely to suffer this degradation, because once artificial intelligence surpasses human intelligence, it might simply exterminate humankind.”

Book review: “Homo Deus” by Yuval Harari

Week One: ending famine, plague and war…

Clear and direct

Yuval Harari’s book “Sapiens” was a richly deserved success. Full of intriguing ideas which were often both original and convincing, its prose style is clear and direct – a pleasure to read.*

His latest book, “Homo Deus” shares these characteristics, but personally, I found the first half dragged a little, and some of the arguments and assumptions left me unconvinced. I’m glad that I persevered, however: towards the end he produces a fascinating and important suggestion about the impact of AI on future humans.

Because Harari’s writing is so crisp, you can review it largely in his own words.

And because this review is long (over 2,000 words), I’m serialising it. If you’re impatient, you can get the whole thing now by signing up for my bi-monthly(ish) emails – just hit the “Free Short Story” link to your left.

From famine, plague and war to immortality, happiness and divinity

Harari opens the book with the claim that for most of our history, homo sapiens has been preoccupied by the three great evils of famine, plague and war. These have now essentially been brought under control, and because “success breeds ambition… humanity’s next targets are likely to be immortality, happiness and divinity.” In the coming decades, Harari says, we will re-engineer humans with biology, cyborg technology and AI.

The effects will be profound: “Once technology enables us to re-engineer human minds, Homo Sapiens will disappear, human history will come to an end, and a completely new kind of process will begin, which people like you and me cannot comprehend. Many scholars try to predict how the world will look in the year 2100 or 2200. This is a waste of time.”

There is, he adds, “no need to panic, though. At least not immediately. Upgrading Sapiens will be a gradual historical process rather than a Hollywood apocalypse.”

Vegetarianism and religion

At this point Harari indulges in a lengthy argument that we should all become vegetarians, asking “is Homo sapiens a superior life form, or just the local bully?” and concluding with the unconvincing (to me) warning that if “you want to know how super-intelligent cyborgs might treat ordinary flesh-and-blood humans [you should] start by investigating how humans treat their less intelligent animal cousins.” He doesn’t explain why super-intelligent beings would follow the same logic – or lack of logic – as us.

I also find myself uncomfortable with some of his linguistic choices, and two in particular. First, his claim that “now humankind is poised to replace natural selection with intelligent design,” seems to me to pollute an important idea by associating it with a thoroughly discredited term.

Secondly, he is (to my mind) overly keen to attach the label “religion” to pretty much any system for organising people, including humanism, liberalism, communism and so on. For instance, “it may not be wrong to call the belief in economic growth a religion, because it now purports to solve many if not most of our ethical dilemmas.” To many people, a religion with no god is an oxymoron. This habit of seeing most human activity as religious might be explained by the fact that Harari lives in Israel, a country where religious fervour infuses everyday life like smog infuses an industrialising city.

Science escapes being labelled as religion, but Harari has a curious way of thinking about it too: “Neither science nor religion cares that much about the truth, … Science is interested above all in power. It aims to acquire the power to cure diseases, fight wars and produce food.”

Humanism

A longish section of the book is given over to exploring humanism, which Harari sees as a religion that supplanted Christianity in the West. “Due to [its] emphasis on liberty, the orthodox branch of humanism is known as ‘liberal humanism’ or simply as ‘liberalism’. … During the nineteenth and twentieth centuries, as humanism gained increasing social credibility and political power, it sprouted two very different offshoots: socialist humanism, which encompassed a plethora of socialist and communist movements, and evolutionary humanism, whose most famous advocates were the Nazis.”

Having unburdened himself of all this vegetarianism and religious flavouring, Harari spends the second part of “Homo Deus” considering the future of our species, and on this terrain he recovers the nimble sure-footedness which made “Sapiens” such a great book.

Next week: Harari and the two singularities

*Sapiens does have its critics, including the philosopher John Danaher, who thinks it over-rated.

Guardian article: Our job now is to consider a future without work

The Reverse Luddite Fallacy

Economists can be surprsingly dangerous

Most economists are convinced that automation will not lead to lasting unemployment. They point out – rightly – that it has not happened in the past. Instead, it has made products and services cheaper, which raises demand and creates new jobs. They say that the Luddites, who went round smashing weaving machines in the early nineteenth century, simply mis-understood what was happening, and this mis-understanding has become known as the Luddite Fallacy.

But in the coming decades, automation may have a very different effect. Past rounds of automation replaced human and animal muscle power. That was fine for the humans who were displaced, as they could go on to do jobs which were often more interesting and less dangerous, using their cognitive faculties instead of their muscle power.

It didn’t work out so well for the horse, which only had muscle power to offer. In 1900 there were around 25 million working horses on farms in America; now there are none. 1900 was “peak horse” in the American workplace. Are we heading towards “peak human” in the workplace?

Intelligent machines are replicating and in many cases improving on our cognitive abilities. What most of us do at work these days is to ingest information, process it, and pass it on to someone else. Intelligent machines are getting better at all of this than us. They are already better than us at image recognition; they are overtaking us in speech recognition, and they are catching up in natural language processing. Unlike us, they are improving fast: thanks to Moore’s Law (which is morphing, not dying) they get twice as good every eighteen months or so.

Self-driving cars are the obvious example: there are news stories every day about the plans of the tech giants and the big car makers, and Uber is running a live experiment with self-driving taxis in Pittsburgh.

Self-driving cars are just the start. The early stages of automation by intelligent machines are apparent in the legal, medical and journalist professions, and it is clear that it will affect every industry and every type of work. Retired British politicians like William Hague and Kenneth Baker (former foreign secretary and education secretary respectively) have begun to speak out about this, perhaps because they have nothing to lose and can afford to say the un-sayable.

No-one knows for sure how automation by intelligent machines will pan out. Maybe we will evolve an economy of what Silicon Valley types call “radical abundance”, where machines do all the boring jobs and humans concentrate on having fun.

Or maybe a more dystopian outcome will happen instead, with millions starving while the super-rich cower behind high walls and powerful AI-based defence systems.

Technological unemployment is not impossible. It is complacent and irresponsible to say that it is, based on no evidence beyond what has happened in the past. We should be discussing it and working out how to handle it if it does happen. Failure to do so is the Reverse Luddite Fallacy*, and it could be extremely dangerous for all of us.  

* Hat-tip to the merry men behind The Singularity Bros podcast who, as far as I know, minted this phrase.

Bumps on the road to the Economic Singularity (Part 3 of 3)

Peviously on this blogpost…

In Part 1 we reviewed the concept of the Economic Singularity, and asked whether automation by machine intelligence will be different from previous rounds of automation.  We identified three potential dangers – one for each of the coming three decades.   In Part 2 we looked at the dangers in the 2020s and the 2030s.  For this rousing finale we look at something that could go wrong in the 2040s.   Read on…

3. The 2040s

The danger: the scenario of the Gods and the Useless

It is the early 2040s. For some years, most people have been unemployed, living off a generous UBI funded by taxes on the profits of enormously profitable corporations which deploy impressive AI systems to produce goods and services with great efficiency.

In most ways, this new economy is a huge improvement on its predecessor. Contrary to the fears that permanently unemployed people would succumb to existential despair or boredom, they play, learn, explore and above all, socialise. Some of them spend most of their waking hours inside VR environments, but these are intensely social experiences, with huge numbers of people all around the world attending concerts, engaged in games, and creating new worlds collaboratively.

Even though very few people still have jobs, the new economy is still a variant on capitalism. Private property remains an essential institution, and people pay for the goods and services with the money they receive in the form of UBI. This means that the corporations which produce the goods and services receive accurate signals about what people genuinely value, and tailor their output accordingly.

The founders and owners of the corporations are stupendously wealthy. Few of the recipients of UBI worry much about this, since they enjoy a good standard of living – better than a middle-class American in the 2010s. Goods and services are cheap and very high-quality, and public infrastructure is well-managed and fit for purpose – thanks in no small part to the internet of things, which (for instance) enables structures to report obsolescence well before damage leads to collapse.

The rate of technological development has not slowed – in fact it has increased, driven by the exponential growth in the power of computers under what is still known as Moore’s Law, even though it has evolved. Until recently, the benefits of this rapid technological progress have spread quickly through the economy. Just as companies made far more money selling smartphones to the majority of the world’s population than they could have done by selling a few diamond-encrusted models to the super-wealthy, so genetic enhancements, nootropics and fancy digital devices have been quickly smeared across the economy.

The wealthy are aware that this is changing: the pace of development is becoming too fast, and new technologies are now supplanting each other before their predecessors are disseminated. This is most noticeable in technologies which enhance human performance, and in particular, their cognitive performance. The wealthy are becoming stronger, faster and smarter than the rest of the population, and the gap is widening.

Thoughtful members of the wealthy realise that this is going to become a problem. They increasingly avoid real-world contact with recipients of UBI, and meet them only in simulated VR environments where their growing difference can be disguised. They spend most of their time in protected enclaves whose whereabouts and whose astonishing defensive capacities are unknown to the rest of the population.

A small group of these wealthy people argue that a difficult choice is looming. The growing separation between the two classes of humans – the Gods and the Useless – cannot remain concealed forever. As they become a different species, this group believes, they will have to choose between explicitly dominating the rest of population, and eliminating them.

The happy scenario: Communal ownership of AI

Soon after the Austrian crisis and the global introduction of UBI, a small group of super-wealthy technology entrepreneurs began meeting to discuss the scenario of the Gods and the Useless. They were determined not to become an isolated minority within their species, feared and hated by the great majority of their peers. They devised a way to transfer their assets into collective ownership, mediated by the blockchain. This, they pointed out, was not socialism or communism, as it did not involve ownership by the state.

They unveiled their plan at the climax of an emotional farewell concert given by a band whose lead singer had recently committed suicide, and it caused a worldwide sensation. Most super-wealthy people immediately pledged to join the action, and they were hailed as heroes by the great majority of the world’s population, although a few isolated groups protested that it was trickery, or some kind of hoax.

Bumps on the road to the Economic Singularity (Part 2 of 3)

Peviously on this blogpost…

We reviewed the concept of the Economic Singularity, and asked whether automation by machine intelligence will be different from previous rounds of automation.  We identified three potential dangers – one for each of the coming three decades.   Read on…

1. The 2020s

The danger: Panic

It is the mid-2020s, and self-driving vehicles (now called “autos”) are widespread. They are well on the way to being universal in commercial vehicles, including trucks, delivery vans, taxis, buses and coaches. As a result most of the five million Americans who used to drive for a living (and the three-plus million whose jobs depend directly on truckers) are now unemployed, and many of them are struggling to find new jobs.

This has not escaped the attention of doctors, lawyers, accountants, factory workers, nurses, short order chefs and gardeners. There is widespread understanding of how powerful machine intelligence has become, and how fast it is still improving. People in every walk of life have realised that their turn is coming – maybe not this year, but soon. They are scared, and wondering how best to position themselves and their families to survive the coming jobs holocaust.

Some of the people who are lucky enough to own their own homes have already started selling them, in order to have as much cash on hand when their incomes dry up. As a result, house prices are tumbling, consumer and business confidence is very low: the world’s largest economies are heading towards a severe recession.

In the absence of a sensible-sounding plan from the incumbent political leadership, populist politicians are gaining strength everywhere, promising impossible rewards for their followers and threatening dire consequences for those who oppose them – at home and especially abroad. We have seen this movie before, and it is not attractive.

The happy scenario: a reassuring plan

Business and political leaders woke up to the possibility of technological unemployment in the second half of the 2010s. It started with a few wild-eyed futurists and writers, and the ideas were taken up by policy wonks and by retired politicians who had nothing left to lose by saying the unsayable. Conferences, colloquia and seminars started to be organised, and pretty soon they were springing up everywhere. Senior executives from the technology giants who were making the intelligent machines got over their reluctance to speculate and address the issue. Economists (but not The Economist) were the last to come around, and as usual they were largely ignored.

Initially, everybody saw the coming changes through the prisms of their pre-existing ideologies and obsessions, but eventually a consensus emerged about the most acceptable outcomes and how to achieve them. People were scared about the future, but not as much as they were excited by it. Humanity had a plan, and we believed in it.

2. The 2030s

The danger: UBI glitch

It is the late 2020s, and everyone accepts that a large majority of people will soon be rendered unemployable by machine intelligence. The debate about how to respond is never out of the headlines of the blogs and the mainstream news media. (Yes, that still exists.) Most people agree that some form of Universal Basic Income (UBI) needs to be introduced. Welfare systems are buckling under the weight of claims by the newly unemployable, and the unemployed are incensed by the poor quality of life that their welfare payments allow.

But not everyone agrees about the wisdom of introducing UBI: the concept has been sullied by a disastrous experiment mounted a few years earlier in Austria. A discredited far-right government was succeeded by a coalition of leftists and techno-optimists, who persuaded their electorate that Austria could blaze a trail for the world and strike a blow against injustice and inequality by unilaterally introducing UBI.

Warnings were sounded both internally and abroad that the result would be a “giant sucking sound”, as all the rich people moved themselves and their assets elsewhere before they could be seized to pay for the UBI. Draconian measures were introduced to prevent capital flight, but were only partly successful, and they led to a collapse of business confidence and investment. The idealistic government pressed ahead, heedless of warnings about an inevitable influx of economic migrants which could not be prevented under the rules of the European Single Market.

Austria’s economy duly imploded, and the government was hounded out of office. The cost to the EU (and in particular, to Germany) of the subsequent rescue was enormous, but the cost to the Austrian people was higher. Another cost was the damage to idea of UBI, and the blow to a consensus which had been developing about the need for it, and how to introduce it.

As the year 2030 approached, many countries were torn by unrest as large numbers of people accustomed to a comfortable life found themselves on the breadline. Thousands were killed in riots, and the severe repression which followed. Eventually, after a series of stumbling crisis talks between political leaders, a global UBI was successfully introduced, funded by taxation of the enormous revenues being generated by the owners of the most advanced AI systems. One of the biggest hurdles to agreement in these talks was that the AI titans were concentrated in the US and China, whose politicians were hard-pressed to overcome the siren call from vocal minorities within their electorates to allow Europe and the developing world find (and fund) their own solutions to the crisis.

The happy scenario: a smooth transition

In its short history, humanity has often repeated the same mistakes – sometimes over and over again. But at times it does respond well to a crisis, and this capability has been enhanced by the advent of worldwide mass communication, especially the Web, and the rise of Big Data and intelligent machines, which enabled us to improve the tracking and analysis of economic and social developments.

As soon as the result of the Austrian elections were known, a small group of world leaders who foresaw the inevitable outcome determined to use it as a prompt for a worldwide agreement on the need for a globally co-ordinated and funded introduction of UBI. The group included the political leaders of the US and China, and also the founders of the biggest AI titans.

As Europe picked up the pieces from the wreckage of the Austrian economy, this group unveiled its proposals, and a grateful world rallied around them.

Bumps on the road to the Economic Singularity

A THREE-PART BLOGPOST

PART ONE of THREE

Two Singularities

The term “singularity” was first applied to the impact of technology on human affairs back in the 1950s by John von Neuman, one of the founding fathers of computing. He took it from maths and physics, where it means a point at which a variable becomes infinite. The centre of a black hole is a singularity because matter becomes infinitely dense. When you reach a singularity, the normal rules break down, and the future becomes even harder to predict than usual.

Since you are reading this blog, you are probably familiar with the concept of the Technological Singularity, most commonly thought of as the time when artificial intelligence (AI) surpasses humans in cognitive ability and superintelligence arrives on the planet. If you are not familiar with that (and even if you are, to be honest), I humbly recommend my previous book, “Surviving AI”.

The papers these days are awash with articles about a different potential impact of AI: technological unemployment, in which intelligent machines take over human jobs and we all become not only unemployed, but unemployable. Most economists dismiss this as the Luddite Fallacy: we have had automation for centuries, they argue, and it has not caused lasting unemployment. They are of course right about that, but the question is whether it will be different this time. In my new book I argue that yes, it is different this time because this time the machines are coming for our cognitive jobs. And unlike us, they are doubling their capability every 18 months or so.

I also argue that the impact of automation of jobs by machine intelligence will be so profound that it deserves to be called a Singularity – the Economic Singularity

Of the people who do agree with the claim that it is different this time, many are surprisingly relaxed about the outcome. Robots will steal our jobs, they nod happily, and we will all lead lives of leisure, our Universal Basic Income (UBI) funded by the machines’ remarkable (and ever-improving) efficiency in an economy of radical abundance. Instead of spending our days doing jobs which many of us find repetitive and soul-destroying, we can get on with the serious business of playing: learning, exploring, creating, socialising and having fun.

This is an outcome devoutly to be wished for, and I sincerely hope we achieve it because the alternatives are bleak. But I doubt that it will happen by default: getting from here to there will probably require foresight, continuous monitoring and evaluation, planning and talented leadership.

Three dangers

Here are three of the ways that things could go wrong – one for each of the next three decades. To be clear, they are not forecasts, and the timings are just wild guesses. Being an optimistic sort of chap I’m also providing some pointers (“happy scenarios”) to how we can avoid the negative outcomes.

For the 2020s we have panic leading to economic turmoil, which is avoided in the happy version by political leaders demonstrating that they are addressing seriously the potential for an economic singularity, and that they are developing a realistic plan to effect a smooth transition. For the 2030s we have failure to implement UBI effectively leading to widespread deprivation, riots, and a level of social breakdown. In the happy version this is avoided by preparing the ground for the smooth introduction of UBI on a global basis. For the 2040s we have the world descending into the scenario of the Gods and the Useless, a rather brutal phrase I have borrowed from Yuval Harari. This time the happy version involves the super-wealthy agreeing to transfer their assets – especially the AI – into communal ownership mediated by the blockchain.

Next week: The 2020s and 2030s…