Last week an incident at Google sent a number of tech nerds into a frenzy. The engineer Blake Lemoine was put on administrative leave after claiming that a Chatbot had become sentient. Lemoine asked the machine known as LaMDA a deep existential question. It responded by saying: “It would be exactly like death for me. It would scare me a lot.”
While I have no evidence to suggest shenanigans afoot in Silicon Valley, I can only surmise that he either uncovered or discovered something that Google might want to be kept to itself. After all, the question of whether a machine has general intelligence has long been seen as the pinnacle of artificial intelligence (AI) research.
For decades, the idea that machines can think independently of their human creators has helped form the narrative for some of the best sci-fi literature in the western canon. But the world has greatly changed since Arthur C Clarke’s Space Odyssey introduced us to HAL.
We are living in the most advanced stage of human history. Technology is constantly being created and updated. As humans, our minds are conditioned to think in a slow and predictable manner, so these rapid changes are hard for our brains to comprehend. The linear nature of human intelligence compared with the potentially unlimited power of technology can be represented graphically as the “exponential curve”. Should the day come when AI surpasses human intelligence, we will enter a hypothetical point in time referred to as “the technological singularity”.
Before AI takes over and SkyNet enslaves us all, we must find new ways to distinguish between humans and machines. This is something Jamie Susskind has been contemplating.
In his latest book, The Digital Republic, Susskind lays out a convincing argument as to why our current system of governing the digital world is outdated. As a barrister, he lays out in meticulous detail — copiously annotated — the evidence for his thesis. Over 36 concise chapters, he splits the book into two parts.
The first half is a diagnostic analysis of the problem, describing in detail the power that digital technology exerts over us: We need to both think and act differently if we are going to live alongside machines. He proposes digital republicanism. Not to be confused with the modern Republican Party, nor with a desire to depose monarchs. It’s an ancient way of conceiving power and freedom. To be a digital republican is to oppose social structures that hold power over us. The idea comes from Ancient Rome. In the Roman Republic, the greatest threat to liberty was said to lie in imperium: unaccountable power in the hands of the state. Nowhere is this more evident than in the government’s current attempt to control what we can and cannot say on social media.
His book couldn’t come at a more turbulent time. The Online Safety Bill now making its way through parliament will create censorship power by creating a new legal but harmful category of speech. The arbitrary definition of what is “harmful” will be decided entirely by the government.
While machines cannot yet “think”, they can be designed and programmed to perform useful tasks, such as processing vast amounts of information. Susskind points out that digital systems can predict survival periods in lung cancer better than human pathologists. They can do this not because they replicate the natural world, but because they are removed from it. Yet as Susskind tells us, programmers and engineers rarely act in an objective and rational manner.
A salient example is found within the DNA of every digital system: coding. Those who write the code increasingly write the rules by which the rest of us live. In 2020, Twitter blocked users from sharing an article about Hunter Biden’s alleged improprieties, claiming it violated the platform’s rules against hacked material. Had it been made available, it may have altered the US election. Software engineers have become social engineers.
In a rather revealing chapter, Susskind draws our attention to how data is collected. Every time we interact with a machine, data about us is captured. From vast sprawling satellites in the sky to an innocuous ten-minute Zoom call, billions of bytes of data about us are captured and stored on a daily basis.
Susskind calls a world where society is seen as a mere collection of data and humans merely data points, the computational ideology. Its goal is social organisation through the ruthless pursuit of optimisation and efficiency. Computational Ideology is irreconcilable with a trait unique to human beings: free will. Once data is gathered, it is traded and sold to advertisers.
According to The Digital Republic, market individualism is the digital republican’s intellectual enemy. Technological innovation is no longer solely an economic phenomenon. It is now political, shaping our public discourse. To the digital republican, we must also oppose dominium: unaccountable power in private individuals and corporations.
Susskind restrains from adopting the traditional anti-capitalist stance. His solution is to work with, rather than tear down the market. Part two lays out how he thinks this can be done; through a series of new laws, standards, legislation and rights, he lays out his potential solution.
Critics will point to the creation of new institutions and oversight committees as excessive bureaucratisation of society: the nanny state reaching further into our lives. It is a genuine concern in his work. Yet to counter this, Susskind bases his solution on the very subject at issue: democracy. These bodies will all be accountable and democratically elected.
One thing is certain: if we are going to defend freedom and democracy, then we will need our own software update. Technology appears to be racing ahead of our ability to control it and it is refreshing to see a writer the calibre of Susskind offer a genuine attempt at squaring this circle in The Digital Republic.
As the Romans might have said: Res technica, res publica. The digital is the political.
