Will AI Take Over The World?

Philosophers amongst you will be familiar with the work of Rene Descartes – a mathematician, epistemologist, and rationalist – much of his work laid the ground for modern philosophy and in particular the strand that has grown out of Hobbes and Locke that informs a lot of the 17^th century and the formation of states and societies thereafter.

There is one eerie and unsettling aspect of his life that is gaining greater attention. Descartes had a relationship with a servant (Helen van der Strom), and their relationship produced a young daughter Francine, to whom Descartes was very attached. Tragically, Francine died of scarlet fever, aged five, and so distraught was Descartes that he had a robot or automata (clockwork, lifelike doll) built in her likeness.

He transported this ‘doll’ with him whenever he travelled (in a casket), and on a trip to visit Queen Christina of Sweden, the crew of the ship on which he was travelling became so alarmed (it was a stormy night) by the robot and Descartes murmurings with it, that they invaded his quarters, seized and broke the ‘doll’ and threw it overboard. Descartes was further traumatized, and whilst it is not clear the incident immediately impacted his health, he died soon after.

Technology scares us

Descartes ‘doll’ is enjoying renewed attention for what it suggests about the relations between humans and machines, how robots can potentially replace and even supplant humans in different ways and for the manner in which this can cause consternation.

The relationship between human and machine is a theme that will cut through the advance (or decline) of the world, and we have written about it frequently (i.e. ‘Talos’). As my limited vision can perceive, will attempt a classification that says there are at least two aspects of this megatrend – the risks that machines take over our (human) world (AI), and the risks that machine led worlds start to exist outside the human one (Defi, Web3/metaverse).

The bad news is that in the case of the former, there is an unknown risk that machines could injure the human race (weaponized AI, the use of AI by ‘bad’ humans and the use of robots in war not to mention the creation of chemical and biological weapons by AI that I referred to in ‘The Final Problem’).

Web3

The good news is that ethereal new worlds – Web3 and Defi (decentralized finance), whose architects had boldly proclaimed were independent of the ‘old’ system, now look like they will be adjuncts to it.

While much of the early hype around Web3/metaverse suggested it was a place that humans could stay in for considerable amounts of time, it now looks like a country they can visit or ‘pop in to’. This much was made clear to me as I attended the Validify digital retail conference in Hertfordshire last week, where the consensus view is that Web3 can help consumers (try clothes or mock up house decors) but will not necessarily become a domain that rivals ‘our world’.

Much the same is true of decentralized finance, which has so far failed to rival in the incumbent financial system, but where its most useful elements such as digital asset infrastructure, are being adopted by incumbent financial system players.

In both cases the growing modesty of new ‘inventions’ is correlated with rising interest rates (and falling market liquidity), highlighting that (as with Descartes time in the Dutch Republic in the 1630’s and 1640’s when the Tulip Bubble took place) many triumphs of innovation are to a large extent cheap money in the drag of new technologies.

In some cases, cheap money and – good design/branding – allows new technology led companies to grab market share, to build new supply chains and to generally make consumer life easier (a small number of fintech and consumer platforms do this). What cheap money also does is allow investors and the wider commercial marketplace to believe that ‘new commercial worlds’ (like the Metaverse) can be created and will have corresponding commercial potential as the human world. The tide is going out on this idea.

To some extent, as expectations of the potential of the metaverse and decentralized finance are deflated, investors and analysts should become more circumspect about AI. AI, the metaverse and defi are very different things – though all driven by the same capital markets, venture capitalists and evangelists.

To my own experience, AI is rooted in data regression analysis – which makes me cynical about it given the time I have spent on econometrics. I do think it is different to Web3/metaverse and defi in that AI can potentially operate in and build out both of these ‘worlds’, as well as ours. AI driven computer programing is an example of such a productivity enhancing application.

What potentially makes it interesting and deadly, to my earlier point, is that it can evolve and enhance the way it has been structured by programmers, to the extent that, to quote Descartes ‘it thinks, therefore it is’. That’s something to worry about.