7 June 2019. Voter turnout among the electorate is like a timid gazelle. Sometimes it appears proudly in the glade of democratic decision-making, while at other times it remains hidden in the political undergrowth. Statistics show: in the past few decades, voter turnout among Swiss people has fallen notably for parliamentary elections. Whilst it was still almost 57 per cent in 1991, by 1995 it had fallen to 42.2 per cent, only to then rise a little again. In general, the statistics shine a light on voting behaviour: political participation is a fickle beast. Some topics bring greater mobilisation than others, cantons sometimes show different levels of engagement and, depending on the proposal, there are also fundamental differences between urban and rural populations.

Is this really adequate in a direct democracy today? Is it appropriate in a time when algorithmic data analysis and artificial intelligence enable us to describe and predict human behaviour almost perfectly, right down to the smallest nuance? And does it reflect the attitude to life among the generations Y and iGen, whose daily lives are managed down to the second by means of their smartphones?

As a direct democracy, Switzerland is already ahead of representative democracies. And yet a lot of things move too slowly and appear too detached from reality and from the people themselves, particularly those under 30, who organise their lives on a day-to-day basis via their smartphones. Direct representation of their lives no longer fits with the democratic practice of occasional voting.

Election forecasts via computer
Why don’t we dare to disrupt the political system? Why don’t we dare to use technology to make a far more direct form of political participation possible and do so with far less logistical effort? In his short story "Franchise" (1955), the American professor and science fiction author Isaac Asimov outlined the early version of an "electronic democracy". In the story, American Norman Muller is selected by chance to decide the political fate of his entire country. He is asked questions and his answers are evaluated with the help of the "Multivac" computer, then multiplied to represent the voting preferences of the population as a whole. Muller is proud that through him, the American people are able to exercise "their free, untrammelled franchise". Algorithmic prognostics replace individual voting.

Things would look a little different today, of course, but the technological possibilities of data evaluation have now become much broader than even Isaac Asimov could imagine back in the mid-1950s. For a long time now, analysis of Twitter data, Google trends and other large amounts of digital data have been used to produce pretty accurate forecasts of how people will shop, invest and otherwise behave. Even election outcomes can be predicted. Hence "Univac I", the first commercial computer in the USA, correctly predicted Eisenhower’s landslide victory based on a sample of one per cent of the electorate in 1952, even though most of the surveys put Stevenson in front. In the US presidential election in 2016, almost all the polling institutes put Hillary Clinton ahead, but South African firm "Brandseye" succeeded in predicting a Trump victory. The data company ascertains moods using algorithms to analyse tweets worldwide and thus predicted both the win for Trump and, even before that, the Brexit vote.

Closer to the people
An algorithmic election based on the computing and forecasting capacities of artificial intelligence systems could outline regularly and precisely what the people of Switzerland actually want. This is particularly interesting for a direct democracy. Even in a system like this, representatives are not always able to get through to voters to convey why political participation through voting is important and to get individuals to express their opinion appropriately. Many people suffer from political boreout. At the same time, politics and its formulation through individual participation is far from dull, yet what is tedious is simply the inability of politicians to address issues specifically, tangibly and with an eye on the future.

In a survey by the "Center for the Governance of Change" carried out this spring among 2,500 adults in the UK, Spain, Germany, France, Italy, Ireland and the Netherlands, a quarter of those questioned said it would be better if political decisions were made by artificial intelligence than by politicians. This reflects, for one thing, the loss of trust that institutions and their representatives are currently grappling with, but it also reflects the notion that technologically supported decisions could perhaps be more accurate, more fitting or even fairer.

An outdated political operating system
In many areas, digitisation is making life more direct. For politics too, no amount of old slogans can detract from this development. Permanent algorithmic representation of the political preferences of the Swiss people would make Switzerland a truly digital direct democracy. Anyone not willing to consider this under the pretence that there are too many risks is simply refusing to acknowledge the problem: How long can politicians go on believing they can motivate people when their political operating system is outdated? For direct democracy, an attempt at a continuous evaluation of political preferences represents not a disruption, but rather a system update. Anyone who refuses to see this is simply setting themselves up for error.

Prof. Dr. Miriam Meckel is a tenured professor of corporate communication at the University of St.Gallen and a founding editor of "ada – the platform for digital life and the economy of the future".

Image: Adobe Stock/RS-Studios