By Robert Skidelsky

This is a "linkpost" but there is no link because this text appeared in a non-public email group. Reproduced here with the permission of the author Robert Skidelsky, a UK economist, biographer of Keynes and economic historian (robertskidelsky.com).

Humans and Machines: Heaven or Hell?

The following post is an amalgam of recent talks I have given on some of the ideas in my latest book, The Machine Age. This book has been published so far  in the  UK, USA, and Germany.
_____

I want to tell four  stories about the relationship between humans and machines. Each offers a vision  of both heaven and hell. After that, I have one more story to tell,  which is the scariest of all. 

Four Stories

First and most familiar is the impact of machines on  jobs. 

Automation  has been  fraught  ever since the Luddites,  early 19th century British handloom weavers, started  smashing the  power looms which were destroying  their jobs and the  poet William Blake conjured up a vision of  dark satanic mills sprouting  across England’s ‘green and pleasant land’.

Over forty years, from 1820-1860,  handloom weaving  was  extinguished.  But in the same period the Industrial Revolution took off with spectacular results. In 1820 the population of the  UK was 25m. Today it is 70m. In 1820 real per capita income was about £1000-£1500. Today it is around £30,000. The simultaneous expansion of population and living standards over the last two hundred years -defeating the dire predictions of Malthus - was made possible by machinery. 

The spread of machinery   not only provided   a seemingly endless stream of replacement jobs at higher wages but supported  a growing  population. 

This  whole achievement hinged on the existence of potential jobs to replace  those being lost.  The question today is: canthis continue?

It  is now  claimed  that up to eight million UK workers could be replaced by AI  within the next ten years; perhaps sooner.

Historically  when you automated  something, the people moved on to jobs which hadn’t yet been automated.  But  you have AGI or artificial general intelligence (some call it superintelligence) the situation  is different. AGI can take over all the  new jobs created by the automation of the old jobs. 

The optimists tell us not to be alarmed by this. They foresee a steady  ascent in the quality of jobs as their routine parts are  farmed out to robots, and humans are freed   for    higher value (more creative) work.  
  
Pessimists like Martin Ford and David Susskind argue that  the   new  jobs created  will be fewer in number and worse in quality than the jobs they replace. They paint a picture of ‘lovely jobs at the top, lousy jobs or no jobs for everyone else’. Dystopian films and fictions tell the same story. Their  trajectory is from the satanic factories of Fritz Lang’s METROPOLIS (1927) to the spaceship of  bloated,atrophied humans of Pixar’s  WALL-E (2008).   

So the current debate is stalled between the  tech-enthusiasts who promote AI as ‘enhancing’ human  performance and those who want to slow it down to avoid replacing humans by robots in all walks of life.

However,  there’s another strand to this debate which goes like this: What’s so bad about having to work less, provided  you receive replacement income?Haven’t we all dreamt of having less to do,having more time for fun and games?

This was technology’s promise  as told by John Maynard Keyes in his Economic Possibilities for our Grandchildren in  1930.

What Keynes said in a nutshell  was  that technical  progress  was bringing Paradise within reach of all. He worked out that in three generations-roughly in  a hundred years from when he was writing  - technological progress would give the prospective population of the ‘civilised’ world  a standard of living between four and eight times higher than in the 1920s,  obtainable at a small fraction of its  current work load. Freed of the burden of toil ordinary people   would be able  for   for the first time in human history to live ‘wisely, and agreeably and well’.  Machines would do all the necessary work, making possible a return to Eden, where ‘ neither Adam delved nor Eve span’.

Keynes didn't actually spell out what people would do when they had no work they had to do.

He was a Cambridge don and reading books, listening to music, and  thinking great thoughts in beautiful settings   surrounded by beautiful  friends and pictures,   might strike some  as   an ideal life.

However, it  may be that many of us would find it  rather depressing to be deprived of  their main purpose in life. Even Keynes thought that the prospect of endless leisure   would produce  a collective nervous breakdown.

At any rate, it hasn’t happened. Most  of us  are  nowhere near a  a  15 hour work week, though standards of living in rich countries are 4 or 5 times higher than in 1930.   

So why have hours of work not fallen in line with Keynes’s expectation?

There are five reasons, two of which Keynes  allowed  for and three others which he ignored. 

The two which he allowed for were population growth and wars. These are infallible means of recreating scarcity just when one thinks one is over the hump of necessity. Today the question is: will Malthus or Mars be the first to overcome  Prometheus?

The three obstacles Keynes  didn't allow for were as follows:

The first is human insatiability.He thought human needs might be quite quickly satisfied, but he ignored the phenomenon of relative wants-I want  something because  you’ve  got it -creating a desire  for more and more, fuelled by relentless 24 hours advertising .Advertising was certainly going on in Keyne’s time but not the relentless, minute by minute advertising targeted at all internet users -now the vast majority of the today’s rich country population- almost forcing  continuous shopping on an addicted population.  

Secondly, Keynes ignored jobs as a source of identity,  and joblessness  therefore, even if compensated,  as  a curse to be resisted rather than a blessing to be embraced.

Thirdly, Keynes ignored the question of distribution, and therefore  of power. He assumed that the gains from efficiency would  go to everyone, not just to the few. But there is no automatic mechanism to ensure this, and  since the ascendancy of neoliberal economics in the last forty years, the social mechanisms for  securing real wage growth have weakened or gone into reverse. While some people have  reduced their hours of work because they can afford to, many others are compelled to work  longer than they want to in a desperate effort to hold on to what they have already got.

To sum up this debate: the assertion that AIs will soon be able to outperform  humans in any task humans do raises the disturbing question: what is the value-added of being human? Will humans not become redundant?  The answer that humans uniquely  have a soul or consciousness is unconvincing to  materialists, who believe, with Descartes, that the soul is located  somewhere in the brain. The human mind is only a complicated kind of brain, and there is in principle no obstacle to building artificial  brains with souls. All such mad thoughts, we must understand,  are  being heavily financed by tech oligarchs. 

My view is that if  we want the work and technology story to end happily  we will need to slow down the rate of job destruction, think more carefully of job replacements, think very carefully about replacement incomes for scaled- down jobs, and institute swingeing wealth taxes to pay for them. 

And we might also  ponder this: how many of the technologically enabled  improvements of (say)  the last fifty years,  can we readily imagine doing without? 
 

My second story is about the impact of technology on health.

 ‘A paradigm shift: How AI could be used to predict people’s health issues’. screamed a Guardian headline of a few weeks ago. An AI trained on 5 years of data and 10bn events such as hospital admissions, diagnoses,and deaths,   will be able  to predict the onset of 1000 diseases, allowing doctors to offer ‘more focussed’ screening tests and  preventive medicines. 

The medical dream is that AI will enable us to live  longer and more healthily. It promises to  reverse, at least partly, the ageing process. We are on the brink of a  generation of technologically enabled centenarians.  Surely this is pure gain?  To which the right answer is that it’s the quality of life matters more than the  quantity of years   -’better to die gloriously than live uselessly’ as the proverb has it. Artificially prolonged life would strike many  people as a horror story. Hence the growing movement in favour of assisted dying. 

A third topic  of current discussion concerns the impact of technology on social cohesion and mental health.  

The great benefit  of the social media  is said to be  the  empowerment of ordinary people through unprecedented access to information.  By  overturning  the authority of professional and religious gatekeepers,  they  release  a pent-up  flood of democratic activity, political and creative.

But with this go the  atomised, solipsistic relationships of humans  with the internet  which replace  person to person relationships,  and which lead to the  growth of internet diseases -  alienation, isolation, pornography addiction, and so on. .The    more digital technology promises to free  its users from the constraints of authority, so does    the demand grow for restriction of access and  control  of content to guard against these addictions.  Here we seem to be faced with a dreadful  race  between technology and mental and social breakdown.

A fourth and , related  focus,  is the impact of of AI on politics. 

Social media are routinely said  to   undermine democracy by spreading  disinformation and conspiracy theories. The control of mass media outlets by tech billionaires   Elon Musk, Peter Thiel, Jefff Bezos, Mark Zuckerberg and a few others  threatens the end of free speech as we have known it. Their collusion with Trump’s clampdown on the media carries the symbiosis between money power and politics to a new level. 

The main  thread running through the stories above is  that as  technology is applied to  an ever wider  range of human activities  much more effort will be   needed  to ensure it remains  safe and  healthy. Prominent leaders in the AI field like Stuart Russell, Max Tegmark, and Yuval Noah Harari  have called for pauses in research and deployment to allow time for reflection on the existential risks technology  poses. The ideal  end of such a pause might be a global agreement to ban, or at least slow down, certain types of research or development,  on the ground that it is too dangerous   to allow it to go forward unchecked. 

But here I come to what is to me the most frightening part of my tale.

Weaponising AI

Some of you may know the famous scene at the start of  Kubrick’s 2001:A Space Odyssey , when  one  of our fur-covered ancestors picks up a bone from a skeleton lying on the ground and realises that it can be used to fight off enemies. Having killed the leader of a marauding group, this humanoid throws up the bone in the air in triumph where it transforms before our eyes into a slender spaceship speeding towards Jupiter. 

This scene  is a timely reminder that technology  started off  as a weapon of war, whether for hunting animals or fellow humans.

I don’t wan’t to suggest that all technology has been developed with  a military purpose. The phrase ‘turning swords into ploughshares’ suggests the exact opposite.Also printing certainly didn’t start off as a weapon of war, though rulers quickly  saw its value for  propaganda purposes

But much more  often than not the  creativity of the scientist  has  been channelled towards military purposes. An early example is  Archimedes who got distracted from pure thought  by the command to build  defences for the protection of Syracus, resulting in the invention of the catapult. And we know  about the great physicists and mathematicians  involved in building the atomic bomb.  Over the centuries governments have continually  subsidised inventors  not to produce a better life but to produce  more efficient ways of killing.  

AI as we know it today  was incubated in war and war preparations.The  computer wasn’t  born in  scientific institutes working for the common good   but in   the UK’s  Bletchley Park   and the USA’s  DARPA programmes, the first designed to break Germany’s wartime code, the second  to keep the USA ahead of the Soviet Union in the Cold War.These developments led to the internet,  first developed for the purposes of military intelligence  ARPANET, a US Department of Defence programme. 

All the voice and facial recognition systems,which are now ubiquitous, were first developed to serve military and intelligence needs.  Of course, there have been  civilian spin offs from which we all benefit.  But their military and intelligence deployment has grown in parallel. My fear is that the latest weaponization of technology  won’t allow for civilian spinoffs because there won’t be any civilians left.

Why do I say this? It’s because all the brakes which we might want to put on AI development for good of us all  will be cancelled when it comes to military development, because we must ensure, that our  AI is better than your AI. 

The optimists will say that even countries at war or potentially at  war will still be able to reach agreements  to stop the development and deployment  of weapons which would cripple or destroy them all. They cite the  Geneva Protocol of 1925 banning the use of poison gas,  and the various  non-proliferation and arms control agreements which  have sought, with some success, to limit possession and development of nuclear and chemical and biological weapons.

These were notable achievements in their day. But such weapons  were specific and  identifiable, so their development was subject to inspection and control. However the threat of  AI weaponry is more diffuse, since it  penetrates  nearly every domain of military operation: drones and robots, intelligence surveillance and reconnaissance,cyber warfare, hacking and  disinformation, command and control enhancement.  

In such a world AI research and development is part of the arms race; AI policy   becomes a matter of making  sure that  our  AI development    stays ahead of that of our potential enemies. 

As Britain and Europe  rearm,   a  key source of funding  for AI development will be the  defence and intelligence services.   What kind of AI we want and for what purposes will be determined  by security  requirements.And this will be true of all countries playing the zero-sum game.

The dominant view today is that we have returned to the Cold War situation, or even worse. Fiona Hill, a member of the UK Government’s  government’s Strategic Defence Review, believes the third world war has already started. The Review itself   demands total society mobilisation against the threat of multiple weapons systems which might be deployed by terrorist or malign states. 

In such a world the difference between peace and war breaks down: we are always at war, need to be on guard against often silent and secret threats. And that portends  the end of free speech, free assembly.

Last week our UK Parliament  passed an amendment to the Terrorism Act 2000, adding  the Palestine Action Group to the list of proscribed organisations. I speak with some feeling about this because the  75 year old mother of my daughter in law was arrested and detained overnight for taking part in one such protest. I’m not suggesting that we have reached the full Orwellian state but there is  an Orwellian creep which requires our constant attention.

So am I an optimist or a pessimist? Is the glass half full or half empty? I would describe myself as a neo-Luddite. Technological innovation, I would suggest, has reached the point of diminishing returns. The small incremental values it can still add to human existence is overwhelmed by the threat of destruction it brings. We have striven mightily for thousands of years to get to this point. Now we should be setting about enjoying its fruits, not gearing up for a new bout  of collective destruction.