A voter’s dream

Jake’s heart was racing. He took three deep breaths to calm his anxiety and climbed the final steps to the electors examination room. He knew he did not need to worry : he majored in economics and social policy and spent hours weekly following the presidential campaign. He memorised every detail from the electoral programmes of both candidates and worked out how their proposed policies would likely impact society over the next decades.

Still, he wouldn’t relax until the computer assessed him and outputted his voter score. How exactly this calculation worked was a mystery, to Jake and most electors, but it seemed fair : ignorance of facts about the candidates’ programmes was penalised, and so was underperformance in the dreaded AI simulation game, where players had to enact policy towards their desired social outcome, in a VR simulation of the real world.

Score highly on both metrics, demonstrate you knew policies and also could understand their effect, and your vote will matter highly. Fail and join the majority of “trumpies”, a popular derogatory term for down-weighted electors, whose vote was of negligible consequence.

Their growing number and dissatisfaction was hard to ignore and Jake recalled the original protests, when New Zealand announced it was going away with the universal principle of “one person, one vote”.

The previous decades had seen us move from skepticism towards AI, to reluctantly letting software automate the boring mechanical processes and business decisions underpinning much of the world’s machinery. Enthusiasm for AI then became unstoppable, when large models demonstrated superhuman ability in domains such as painting, literature and debate.

This was also a time of political unease, with increased polarisation, surprising, and to some worrying, democratic outcomes all over the world, from US elections to UK referendum. Some claimed this was a failure of education and democracy. Our people simply weren’t up to the task of making good decisions, it seemed.

AI managing our society and politics was one step too far for most, but letting a machine judge each voter’s aptitude seemed plausible.

Bias was an obvious worry : the proposal to use AI came from the liberal left and was quickly branded as a way to mute the more vocal and politically active conservative right.

Judging a voter’s aptitude needed to be just that : how clearly they could articulate their goals for society, understand how to get there through economic and social policy, and know which candidate would enact the policies best aligned with their interest.

It did not matter what their goal was : every voter was free to imagine the future of their dreams, without moral judgement.