“Hello, Hal, can you help me about AI?” is the second in a series examining the shift in society caused by technology by Theo Wooster. Building on “Clogs & Conspiracy Theories“, the view shifts to the wider impacts of economic shifts and starts to hint at what can be done to share in the benefits.
The notion of a machine capable of thinking and acting for itself, has been a topic of myth, legend and speculation for millennia – from the ancient Greek myth of Talos, the bronze automaton who protected Crete from marauders, to the Tin Man in the Wizard of Oz and beyond.
From progress made over the last decade, out of the thousands of years such machines have been mythologised, we are on the cusp of these becoming reality.
The pace of innovation and development in the fields of machine learning, artificial intelligence and robotics over the past decade has been staggering – the capabilities of machines that can “think” have enhanced at an unprecedented rate since the “AI winter” of the latter half of the 20th Century and early 21st century.
“Narrow AI” is a form of machine intelligence that is already commonplace – able to do specific tasks involving reasoning and problem solving. This form of AI is an area of huge growth in a massive variety of applications for business, manufacturing.
But this time it’s different
However this revolution, unlike those that have gone before, is less concerned with traditional barriers being equally at home in traditionally reserved professions such as medicine and journalism; this time it won’t be just Luddite artisanal labour that is affected.
Artificial general intelligence is the term used to describe machine intelligence that is more comparable to that of a human. Tests to assert whether machine capabilities have reached this level of ability include the famous Turing test and the Wozniak test.
Whereby a machine and a human converse with a second person, and the machine passes the test if it can fool the second person into believing that it is talking to the human subject.
Whereby a machine can enter an average American home, and brew coffee, finding the necessary ingredients and push the correct buttons.
There is speculation that machines may pass such tests within a foreseeable future, with progress being made all the time – for example, last year, Open AI built a system that beat the worlds best DOTA 2 players – a three dimensional, real time strategy game involving teamwork – the latter aspect deemed impossible by many just a few years ago.
The speed at which both forms of intelligence are developing is unprecedented, but how fast is progress really being made in comparison to the development of previously ground-breaking technology, and should we be worried about it?
A Glance to the Past…
Artificial intelligence is undoubtedly a technology with a great potential to change the way society functions, but this is hardly the first technological revolution. Parallels may be drawn with the advent of mass-produced motor cars, personal computers, and automated factory machinery.
A feature of each economic revolution is that it tends to play out more quickly than the previous, and in that regard artificial intelligence (particularly in the current narrow form) is no different; it is being integrated ever more quickly into common usage than any of these previous technologies.
In 2019, a survey by McKinsey found that 58% of businesses had integrated AI into an area of their business, up 11% from 2018. Furthermore, a 2020 Cognilytica survey found that over 90% of businesses are planning to implement at least some form of AI in the short term.
To the speed of the future
In 2019, there was an estimated 3.25 billion voice assistants (e.g. Alexa) worldwide – roughly the same number as global smartphone users that year – and the number is expected to grow to upwards of 8 billion by 2023.
To put this in perspective, the Amazon Echo was launched in 2014, so in the space of 6 years the smart assistant market has grown the match the volume of the smartphone market which arguably launched with the original iPhone in 2007.
The smart assistant sector has grown at twice the rate of even the smartphone market.
As humans, we are remarkably bad at spotting the pitfalls of revolutionary technology (no matter how beneficial) until late in the adoption cycle.
From an automotive standpoint, think seat-belts – the Ford Model T became the first mass-produced car in 1908, yet seatbelts were not mandatory equipment in the USA until 1968, and usage is still not compulsory in all states.
A tragic example, linking back to the previous article, is the consequences of the introduction of automated looms into mills in the 18th and 19th Century.
Not only did many skilled technicians and artisans lose their livelihoods with the introduction of machine looms, it sadly also resulted in a greater use of child labour in mills as low-cost 15 year olds became able to produce more textiles than more expensive artisans using more traditional techniques.
A clear trend emerges that before any major economic restructuring, we, as a society, are wholly and repeatedly unprepared.
Society was similarly ill-prepared for the industrial revolution, the shift to a service economy with the closure of mines, or the current trend towards “thinking” technology but each transition point has typically led to wider societal impacts.
For instance, take the UK’s transition away from manufacturing and the primary/secondary sectors towards services and the tertiary sector from the late 1970s onwards.
In contrast to the introduction of mechanised looms in the mills which, while significant for those affected, had a relatively narrow negative impacts on wider society, the transition which started in the ‘70’s left many sectors whose production was no longer considered necessary, together with their often closely knit communities, largely redundant.
Many of these communities remain without the necessary skills and infrastructure to integrate into the “new” service economy, leaving them deprived and isolated, and contributing to the rising inequality in the country.
As discussed in the previous article, the impact of smart technology and artificial intelligence is set to leave few sectors untouched and, for the first time, will directly impact the professional classes.
With so many affected, the impact will be almost universally felt if we fail to prepare as we have failed in the past.
“I’m sorry Dave, I can’t do that”
In fact, the only people who appear immune to the AI restructuring are those fluent in the technology and the mathematics that fuel it.
“Hello, Hal, can you help me with AI?”
“I’m sorry, Dave, I can’t do that”
Introducing artificial intelligence, unlike automatic looms and lorries, requires relatively low levels of capital and investment for remarkably high rewards – the only prerequisite is having the necessary skills and cognitive ability to create and manage it.
Hinting at a Future
One thing that the COVID-19 lockdown has demonstrated is that already huge proportions of our now largely service focussed economy are able to work from home.
With near universal broadband, this means anyone with the requisite skills, wherever they live, can become an AI entrepreneur, technician or developer – all that’s needed is the relevant training.
This presents a unique opportunity to re-level an economy which has been increasing inequality if the right steps are taken – a child from an ex-mining village in County Durham has the same raw ability to learn mathematics and computing as an Eton schoolboy.
AI presents so many opportunities and positive repercussions for society, from improving living standards and education to increasing productivity and environmental conservation.
But it also carries with it the potential for significant fallout and structural unemployment. The speed of the arrival of smart technologies means we don’t have 50 to 100 years to consider the impact first hand and to develop evidence-based policies.
If society wishes to capture the gamut of opportunities and avoid the long-term damage and inequality that has followed previous economic revolutions, action needs to be taken now – rather than sitting idle and allowing the pod-bay doors to close in front of us.
The third part of this essay will start to look at what can be done to ensure society enjoys the upsides, while mitigating the downsides of the AI Revolution.