Technological Cliffs
A.I., idols and the future of human life
AI is constantly in the news right now, whether its Sam Altman (OpenAI or ChatGPT’s founder) appearing before a bipartisan Congress to push for regulation, or public estimations of job losses — just on Thursday BT announced job cuts to some 20% of its workforce and last month Forbes estimated some 300 million worldwide jobs, that’s roughly 10% of the global workforce, would have their jobs replaced by AI. Others are more optimistic thinking a productivity boom and new jobs will quickly fill the old, as previous technological revolutions have done.
So in part, then we are living through an age-old question — how does society respond to rapid moments of technological change? To what extent is the AI revolution a simple evolution of the long industrialisation process (mechanisation, mass production and automation) and to what extent does it represent something deeper that should be of concern to us all? When I was prepping for this seminar I had expected to do a relatively straight down-the-line look at the adoption of technology by society and the positive and negative trade-offs that that brings with a theological consideration of what it means to be human amongst this. Whilst that is part of this seminar the concerns raised by AI researchers (of all stripes) and Christian scientists too meant I ended up down a surprising path in Revelation that I hadn’t bargained for. This latter bit is speculative so will need to be treated with appropriate caution, as all treatment of apocalyptic literature should be, but the scope of what AI represents means the discussion is inevitable.
So, in terms of the scope of this we’ll consider past technological shocks and the impacts this had on society. Secondly we’ll go through a direct consideration of what is occurring now and how well we are, or are not adapting to this. This is all the envisaged straight down-the-line material and is the least speculative. Thirdly and finally, we’ll take a look at the ideology underpinning many of the hopes for AI silicon superintelligence, and the capacity it may represent — much of which is clearly antithetical to Christian ethics, and Christ altogether. So Past, Present and Future and we’ll start with the easiest which is the past.
The Luddites
On March the 11th, 1811, a group of textile workers broke into the local textile factory, where out of fear of automated loom machinery and the threat to their way of life they decided to fight back proceeding to break up, or set fire to the hated machinery. These were the original luddites and this act of vandalism became, to some extent unfairly, associated with any scepticism of technological advances. However, technological advances should always be treated with caution, not because there isn’t good to be welcomed but because the unintended consequences can be far reaching.
For the luddites, their jobs were largely replaced with better ones but their way of life was decimated and the rise of an urban class did lead to a progressive hollowing out of rural life and culture — the advance was double-edged. This is feature, not a bug of change, and technological advances usually bring significant cultural and societal changes, some of which may be predictable but many of which will be unintended and harmful. That doesn’t mean we shouldn’t accept change, but also means yelling at the obscurantists misses what it is being lost.
The Past
By way of further example I want to look at three further major technological changes in the last 500 years and consider their impacts, and then get yourselves to consider a fourth (i) firstly, the Gutenberg printing press — on the one hand this increased the spread of knowledge, helped to standardise language, increased literacy, underpinned the reformation by allowing theologians to print and distribute and facilitated the emergence of a public space — set against this, overproduction and waste became a new normal, and polarisation and extremist responses became a feature — for e.g. the 30 Years War that was in part a conflict between Protestant and Catholic states within central Europe accounted for nearly 8 millions lives — one of the bloodiest wars in Europe’s history. One might say, justifiably, that this war spun out of the cultural changes that occurred too quickly for society to adapt to.
(ii) Secondly the Steam Engine — this was the industrialization Big Bang, and vastly increased productivity by replacing manual labour with powerful repeatable machines that didn’t need to sleep, the transportation of goods and people over long distances, and the subsequent the rise of an urban middle class reducing levels of poverty and the growth of leisure and commerce — set against this of was the hollowing out of rural life and the long-term environmental impact from destruction of habitats and deforestation, carrying implications through to the present day.
and (iii) Thirdly electricity, leading to new forms of communication, of learning, the manufacture of white goods transforming domestic life, further automation of machine labour and further increases in productivity, followed by the rise of a globalised culture through international trade and exchange — set against this of course the mass use of electricity required fossil fuels further impacting the present climate crisis, consumerism reduced social ties and separated people from local communities leading to increased atomisation and loneliness and the now to the current spread of surveillance reducing privacy and autonomy.
I hope by now you see a pattern — improvements in productivity and living standards come with trade offs that are far-reaching and often with cultural changes that can be deeply unsettling and atomising.
The Near Present
Has anyone here used ChatGPT? I have to confess I have enjoyed and been surprised by its capacity — its ability to summarise and recall an internets worth of information and reply with human-matching eloquence. This is the new era of silicon based intelligence but isn’t the same as consciousness and it’s important to make that distinction. So with respect to the present A.I. what exactly do we mean?
‘Machine learning algorithms capable of continuous improvement towards a defined goal.’
Let’s consider this briefly through an example. I am, as some of you may know, a big fan of chess -when Gary Kasparov lost to Deep Blue in 1997 he lost to a powerful, but largely static calculator. This same calculator had no capacity to improve with time and would have lost to a 4 year old at noughts and crosses. It was an impressive feat of computational engineering but in a very narrow domain. The modern equivalent is MuZero a neural machine-learning algorithm that vastly exceeds DeepBlue and is self taught through pattern recognition and has the ability not only to improve with time but also to master a broad suite of games including the more complicated Go. It is an example of silicon based intelligence that vastly exceeds human capacity across a now wider domain, with a simple goal in this case of mastering games.
Algorithms such as these now impact most of our lives in increasingly sophisticated ways and in different political spheres. In the West they run much of a consumerist or capitalist system, and whether through your laptop or through Alexa are essentially data collectors that we voluntarily allow into our home to monitor our likes and dislikes (surveillance capitalism), fed back to an algorithm that is tailored to get us to spend more (Amazon) or engage more (Social Media), and the algorithms don’t care about whether your spending is sensible or your social media engagement is helpful, etc… — the defined goal of these algorithms is getting you to voluntarily give more of yourself to central systems so they can better tailor goods for sale, or political campaigns for persuasion. In this respect, social media has largely worked against humanity and mutual distrust and dislike across the political sphere has run dangerously close to boiling point across both sides of the pond.
So that’s surveillance capitalism, but what about the East. Well in China, there is now a sophisticated social credit system called Skynet — a universal surveillance system using 700 million CCTV cameras (half of the world’s total) to enable real-time spying of its citizens. Skynet’s use of everything from social credit to facial recognition increasingly requires significant submission to the state or the risk of becoming one of society’s unmentionables. The consequences of which are your details, ID and addresses being splashed across government billboards and having your access to public (i.e. transport) and private goods (i.e. shops, hotels) severely restricted. This form of surveillance communism is a new digital leprosy, for that is how it is enforced, and looks uncomfortably and increasingly dystopian.
Now, in case this is in danger of sounding too gloomy, AI algorithms are in the end tools and also increasingly turned towards noble problem-solving challenges geared towards life enhancement. So, one domain where changes are happening quickly and for the best is in diagnostics. For example AI systems have now been trained to such a level that matches or exceeds experts in the assessment of diabetic retinotherapy, or again the capacity to perform high precision neurosurgery within an MRI — one of the treatments that could be performed to alleviate Parkinsons. Again, AI is already used with radiology to enhance the process of diagnosis. Progress in the application of AI to diagnostics could free up stretched health services across the world, potentially to be re-invested in front line person-to-person care. There’s a whole host of other fields where AI can, and is, capable of vastly increasing productivity including language translators, transport and industry, as todays techno-optimists are keen to point out.
Now as intelligent and sophisticated as these various tools are, we should understand that at this level they are, for the moment, just that, tools, that are currently being used by individuals, corporations and governments to bring about a variety of noble and nightmarish aims. Some of it is manipulation, some of it is Machiavellian, some of it is brilliant and life-enhancing. In that respect this current set of tools can be considered in the same way to any rapid technological change in the past and its impact on society. The pros and cons can be weighed up and regulation can be implemented to offset the negatives where required. Worldwide agreements have been made in the past to prevent possible research as shown by research into Human Cloning which was successfully halted, or Biological Weapons , which are severely curtailed— it is possible, and international agreement may be possible to curtail the worst aspects of A.I.
However, many if not most researchers and thinkers, Christian especially but many others are of the opinion that the speed of A.I. development is vastly outstripping any ethical or regulatory framework that is necessary, and there are reasons to be concerned about some of the hopes being pinned on the new technology.
The Future
“Does God Exist ?”— “I would say not yet” (Ray Kurzweil)
On March 22nd of this year (2023) an open letter was signed by key researchers [1] from across the A.I. space calling for a general pause in the development of A.I. resources. Along with the number of world renowned signatories I was struck by some of the comments around the edges reflecting the ideology of those who haven’t signed and are pushing for further and faster research. Some of these ideologies reflect either a profoundly anti-human disregard, or are looking at A.I. as a potential benevolent God — a true 21st century idol as it were, and both of these are reasons for all of us to be concerned.
So two interviews I found interesting was Elon Musk being interviewed by Bill Maher, and Tucker Carlson [2,3], two wings of American politics and in both interviews he cites two interesting comments, one from Ray Kurzweil who see A.I. as solving the problem of human longevity (that is effective immortality through digital emulation, or, I think, using A.I. to figure out how to reverse / halt physical ageing). And secondarily one from Larry Page (CEO of Google) seeing the desire to protect humanity against a potentially superior silicon based superintelligence as being “speciest”. Both Ray Kurzweil and Larry Page belong to the further, faster strain of A.I. research.
Max Tegmark, one of the original A.I. optimists is now far more circumspect and referenced a famous 2014 essay called “Meditations of Moloch” [4] but even this story, a reference towards humanities capacity for self destruction, and the sacrifice of moral good to achieve temporal success ended with the desire to construct a God-like creature that would prevent this, serve humanity, and kill the worse gods of our desires. This is many researchers hope for superintelligent A.I. and it is something that causes me, at least, significant unease. I simply wouldn’t trust that such an idol, and that is what it would be, would be anything more than harmful in the long run, and particularly so to Christians. There is, as mentioned some very uncomfortable echoes of Revelation in all of this. In particular some of the ideas from Revelation 13 that look towards a World Government where only those who are marked can buy and sell — echoes of which we are beginning to see in miniature.
Now most of this is speculation. Superintelligence isn’t here yet, but the expectation across nearly all current A.I. researchers is that superintelligence will be here one day, and within our lifetimes. Apocalyptic literature is notoriously easy to read badly and Christian communities throughout history have at times believed that their particular phase of history is the last beast of Revelation when at most it was a reflection of the general characteristics of that beast. One can get carried away and I have no desire to make predictions. Nonetheless, there are some general questions that we should make and think about as a matter of urgency, particularly given as to what and how superintelligence — defined here as a computer based artificial intelligence that exceeds human capacity in most, if not all fields, is used.
Researchers have suggested several scenarios but let’s assume the following, all of which are considered likely. (i) humanity creates a superintelligent AI. (ii) it isn’t conscious and remains as a tool, a very powerful tool, under the control of its creators (benevolent God scenario) to be used to improve humanity (iii) that control is being used by a global corporation(s) and / or collective government(s) for, as they see it, the betterment of humanity. The scenario calls to mind C.S.Lewis’s prophetic idea in the Abolition of Man.
“What we call Man’s power over Nature turns out to be a power exercised by some men over other men with Nature as its instrument… Man’s conquest of Nature, if the dreams of some scientific planners are realized means the rule of a few hundreds of men over billions…. Each new power won by man is a power over men as well:” C.S.Lewis — Abolition of Man
Whether he knew how prophetic he was being I don’t know, but AI, if developed as the tool most researchers envisage and without even the more speculative scenarios being publicly discussed could represent a tool that gives very few the power to oversee the lives of many billions.
Final Idea
Huxley and Orwells dystopian imagining of the future considered the future to either be humanity beset by triviality or a future human society under heavy surveillance. The combination of the West’s surveillance capitalism, and the East’s surveillance communism means these two could be sadly combined in an overall society under heavy surveillance living out trivial pursuits. It is a possibility increasingly plausible with a superintelligent AI tool as we can see in the blueprint from China for emerging total digital surveillance and in combination from the West where we may increasingly be encouraged to live out atomised lives seeking dopamine hits in VR. The challenge for Christians would be there would be no easy opt out — and this is at least implied in Revelation literature (see Chapter 13). As Christians we need to think not just about the Pros and Cons of AI but what it means to live in the modern and emerging world as exiles in one form or another — belonging to our families, belonging to our communities, belonging to each other and able to hold fast to Jesus as the true benevolent God, rather than hand salvation over to an AI God of our own making.
References
[1] https://futureoflife.org/open-letter/pause-giant-ai-experiments/
[2] https://www.youtube.com/watch?v=a2ZBEC16yH4&t=599s
[3] https://www.youtube.com/watch?v=oO8w6XcXJUs
[4] https://slatestarcodex.com/2014/07/30/meditations-on-moloch/
