The COVID-19 pandemic has plunged us into scenes that feel like science fiction films. Downtown streets are eerily deserted, shop staff look at us through shiny face visors, we watch vehicles spraying disinfectant across cities on the other side of a planet, and everywhere, there is an unseen enemy.
One sci-fi scenario that has come nearer is immersion in a digital matrix. Under lockdown, our digital life has deepened. We shop even more online, day-to-day meetings are networked, the screen has replaced buildings from museums to cinemas as the stage for shared diversions, etc. The level of our activity that is traceable and can be monitored has jumped. A new form of individualised and collective surveillance has emerged in contact-tracing apps. New rivers of data have started to flow. Technology can overcome physical isolation, but it also amplifies its effects.
I have previously warned that we are sleepwalking into the digital matrix, and have speculated a future scenario which I have called ‘The City of the Singularity’ — a state where superintelligence micro-manages the city and every citizen in it. Superintelligence is artificial intelligence (AI) which exceeds human intelligence, and is reached at a stage called the Singularity, which is regarded as imminent by some experts and impossible by others. In my hypothetical city, I call the AI that runs it an Ubiquitous Super Intelligent Urban Service (USIUS).
Ultimately, a USIUS could make city life indistinguishable from a simulation — at least to us as humans. But we will not be alone in the City of the Singularity, and today’s world already gives a preview of the other characters we will live with. In this blog, I ask:
How do we share society with intelligent digital beings?
Other thinking entities might include trans-humans, who are people upgraded with digital implants which, for example, enhance senses or plug us directly into digital networks. We’re already just a step away with wearable technology, like digital hearing aids or augmented reality glasses. Transhumans and all the issues they raise are yet to come, but synthetic, intelligent workers are already here in the form of onscreen agents and autonomous robots.
When technology does something a human can’t, we benefit. That can range from AI spotting skin cancer in a photographic image, to Collosus, the robot which helped extinguish the Notre Dame fire in 2019. But when technology does merely what humans already do, is there any benefit other than someone’s profit margin? Humans have been replaced by robots on the assembly line. Robots can replicate the lifelong-learnt skills of the artisan. Train drivers, truckers and ship captains are, or soon will be, replaceable. Instead of humans, chatbots already deal with us on commercial websites. Animated computer-generated characters are now surfacing. If a fashion website uses synthesised models to show its garments online, that devalues the human who works as a model. The same goes for virtual assistants created to deal with online customers, such as Soul Machines’ personas. Who needs humans in a call centre or a bank branch when they can do the job? This technology also happens to be just a step away from deep fakes, which go beyond replacing jobs to hijacking identity.
AI is already moving into creative fields. AI can design, sculpt, write, compose, and more. Some artists say AI is merely a tool for them to use, but to a deep-learning set of algorithms, an artist’s output may be merely raw data for them to emulate. In 2016 a computer generated a new Rembrandt painting. If a pop song is written by AI, sung by a synthetic voice and presented by a computer-generated singer, how much human musical talent does it displace from the music market? Where is this going? If AI can do anything and everything a human can, what purpose do we serve?
We need to challenge the digital devaluation of human work and creativity, because they give us purpose.
In 1942, Isaac Asimov wrote his ‘Three Laws of Robotics’, starting with: ‘A robot may not injure a human being or, through inaction, allow a human being to come to harm’. That needs to be extended from robots to cover virtual characters, all algorithm-driven machines and AI systems themselves, including a USIUS. Harm to a human being includes robbing them of their livelihood. Should we have a basic rule for all new digital agents that they only get to be sanctioned if they are job creation-neutral or -positive? It’s not necessarily straightforward to predict. Until now, technological revolutions have tended to expand economies and create more jobs, although not always for the less educated. In 2019, Kai-Fu Lee, an investor with a portfolio in billion-dollar AI companies, predicted that automation would make 40% of jobs ‘displaceable’ by 2035. Ironically, the only way to evaluate the effect on jobs of new AI will likely be with new AI.
The argument that automation liberates people from menial jobs is good if there is more fulfilling work to migrate to and educational resources to acquire new skills and outlooks. Free time is fine if people have financial resources. That’s a big if. Affluent economies could pay everyone a universal basic income. Commentator Aaron Bastani goes further, arguing that technological progress can produce such a high level of (what are now) luxury commodities that ‘everyone has the means to a good life’, provided that politics reject capitalism. He calls it Fully Automated Luxury Communism. This is clearly a new variation of consumerism, and the Society of the Spectacle described by Guy Debord, and it doesn’t address the human need of a sense of purpose.
It is also one of the scenarios that a City of the Singularity delivers. And if no-one at all was working, a superintelligent agency would be needed just to keep things going.
Whatever sort of technological future we have, humans will co-exist with intelligent artefacts. As the AI in the digital agents we share society with tends towards superintelligence, new questions arise. If embedded AI could give a machine in some way a consciousness, does it have rights? If animal rights are an issue, why not robot rights? Furthermore, is there a threshold of intelligence at which a robot could claim citizenship? If we started thinking of granting rights proportional to machine intelligence levels, that would undermine the egalitarian basis of human rights for all.
These questions arise partially because of commonplace ideas of what a robot is. It’s easy to think of them as cyborgs, basically humanoid in form, but we already know that they take very different forms. Contemporary robots like those on an industrial assembly line or autonomous self-driving vehicles look nothing like humans. They’re pretty dumb, but imagine when they become as smart as the thoughtful bomb in John Carpenter’s 1975 film Dark Star.
When I proposed a the USIUS that micro-managed all aspects of the City of the Singularity, including the lives of its citizens, I imagined that this AI could be across a distributed system. Many species achieve swarm intelligence which is at a far higher level than its individuals. Termite colonies, for example, can collectively build incredibly complex architecture to live in. Swarm intelligence is facilitated by a network of communication across the swarm — in the case of termites, chemical signals in pheromones they secrete. We are already building a communication network for digital devices to talk to each other in ‘The Internet of Things’ — it’s called 5G. Could a future generation network enable devices with built-in AI to a swarm superintelligence? Could the foundations of a USIUS be future-generation telecoms network architecture – even the next one, 6G?
While devices connect on digital networks, we have been disconnecting in the real world — long before COVID-19. Loneliness is on the rise, and the Japanese condition of hikikomori is going global. Meanwhile, whole societies are under digital surveillance and AI is feeding on big data to understand us better than we do ourselves. It may use that data to sell us something, protect us — or control us. Soon, we may not be smart enough to think ahead of AI. That is by definition the case if the Singularity happens.
Now is the time to wake up to the issues, rather than sleepwalk into purposelessness as the digital matrix immerses us.
This post dated May 2020. © Herbert Wright
I am indebted to Simon Tyszko for permission to use his images.
A final part of the City of the Singularity blog series will consider the built environment, raise the issue of the Emotional City, and planning for the ideal with or without AI.