Liquid Democracy

Using Technology to Expand Democracy

Awe and empathy don’t easily scale. The awe inspired by the Milky Way on a clear night could not be magnified by 100 billion on a chance sighting of the entire universe. Empathy too has a limit, the keenness of grief felt for one loss of life struck dumb by mass tragedy. And to this human shortcoming we might add democracy.

As India takes to the polls, some 900 million people are casting their vote in the largest democratic exercise in history. This is a population a thousand times greater than that of Ancient Greece, where democracy was first conceived. Yet population growth and globalisation creates a complexity that threatens the workability of democracy: then, as now, it is challenged by the autocracies it sought to curb as they eke an advantage in long-term planning. Far-sighted infrastructural decisions must be taken: this is how (silk) roads are laid and technology developed. Keeping a public informed of every matter is unworkable, creating inevitable ignorance on points that need addressing politically. Of course, part of the value of ‘one person one vote’ is that every member of society can express how the government affects their livelihood, expert or not, but under autocratic rule, enormous costs can be spread over decades, levelling biases of short-term budgets and indecision inherent in democratic cycles. Of course, one hopes for a benevolent dictator, but failing this liquid democracy might offer a solution.

In ‘The Best and the Brightest’, David Halberstam notes the moral compass of President John F. Kennedy that affords him to quickly intuit decisions. Not that facts eluded him, but that there are simply too many to gauge at any one time. Having to act on a hunch is a necessary quality in a president. The dream for AI and machine-learning is that all facts can be identified and analysed, with perfect solutions calculated. There are no bad days.

But the system has some hurdles to clear first. A study last year at University California revealed the algorithmic biases of machine learning. By examining mortgage lending sites, AI was found to be identifying applicants less likely to use comparison sites and offer them higher-priced mortgage products. However the people less likely to shop around were often ethnic minorities, embedding discrimination as part of the tool. “The mode of lending discrimination has shifted from human bias to algorithmic bias,” said study co-author Adair Morse, a finance professor at UC Berkeley’s Haas School of Business. “Even if the people writing the algorithms intend to create a fair system, their programming is having a disparate impact on minority borrowers—in other words, discriminating under the law.”

In this new landscape tilled by AI, liquid democracy emerges as a political shoot of Decentralised Autonomous Organisation, a system that negates central command by devolving responsibility to involved parties. Through this, voters could have their say on a whole swathe of issues, from general elections to referenda. Budgets as a manifestation of political intent can be scrutinised. One voter might favour more money spent on infrastructure; another more in healthcare: liquid democracy enables this difference to be articulated.

The system also allows specialists to have greater say over their expertise and relieve voters of subjects beyond their scope, an opinion that in a traditional democratic system would hold as much sway as an expert. Using blockchain, if you don’t have an opinion on a matter, you can delegate to someone who does, someone who could even delegate to a third party, creating a sequence of trust on a given issue –a process known as transitivity. This principle of integrated expert opinion could help expand democracy from 100 million-strong populations to those with over a billion.

The flaw is that not everyone has the time to weigh up the pro’s and con’s of every scenario, nor the intuition of JFK. Thus the people who have the least time and mental capacity are more likely to delegate their choices, implicitly handing power to involved parties. This could penalise single parents, for example, or those having to hold down several jobs to make ends meet. It also requires a degree of expertise to recognise another expert, and choosing to delegate is another decision the voter would have to face. In this new market of ‘experts’, we might see members of a community vying to be delegates alongside day-time TV hosts, or Instagram influencers appealing to their followers. It is easy to see how this model could be exploited.

In extremis, liquid democracy has the potential to render political parties obsolete by opening every topic to the public and capturing the atomisation of opinion. There would be no need for the electorate to identify with left or right groups; voters would have a say on every matter, a ‘line of best fit’ calculated to govern by that eliminates the need for electoral cycles. But before liquid democracy can be implemented, it will first require a party to enact all that is elected. In the pursuit of democracy we may need to hand the keys to the future to technocrats, meaning it may not be the politics of leaders we vote for, but those of the coders.