Recently, the IE University’s Centre for Government Change asked 2,769 respondents how they would feel about removing some of the politicians in their government and replacing them with artificial intelligence programs.
Despite carefully laying out the “clear and obvious limitations” of AI, 51 per cent of Europeans thought this would be a good idea, with a high of 66 per cent in Spain.
For Chinese respondents, the number climbed to 75 per cent in favour. But in the bastion of democracy south of the border, 60 per cent said they would not be in favour. Strangely enough, despite everything that has happened in the past year, the United States still believes in democracy.
Perhaps more importantly, they believe decisions about humans should be made by human beings. Skynet will not rule the U.S.! Germany and the United Kingdom were also not in favour of letting computers takeover.
What do these results tell us? Oscar Jonsson, director of the centre and one of the researchers contributing to the work, suggests the results are indicative of the general disillusionment which exists in society for democracy.
But given the diversity of respondents from a variety of political system, it is perhaps not so much democracy which is at issue but politicians. When broken down by age, 60 per cent of the 25-34 years old cohort were in favour of AI politicians falling only slightly to 56 per cent in the 34-44 years old range. Not surprisingly, the 55+ crowd were dead set against the idea.
Asked pretty much any question about the present state of politics and it tends to be young adults who are disillusioned by the process. I know that when I was younger, I once voted for the “Rhinoceros Party” because none of the federal candidates in my riding seemed to have any understanding of the issues important to my generation.
Of course, I am older now. And I know that politicians rarely represent the collective voice of an entire riding.
Indeed, it would be difficult for anyone to bring together the collective opinion of a riding. Most politicians find people in their own party – who, in theory, are supposed to hold to the same political values – often disagree with them. MPs often find themselves at odds with the official party position.
But would AI solve any of these problems? And what would they base their decisions on?
Despite the illusion of intelligence provided by computer programs, they are – after all – computer programs. They are programmed algorithms for solving problems. They are often writing in mathematical coding with logic defined by boolean relations.
The coding is complex. No doubt about that. Often programmers are left shaking their head at the things their creations come up with. The structure of the algorithms and the extent of coding a simulated intelligence often makes it so that the programmers really aren’t sure exactly what they results will be. They have a good idea what the program will do but not an exact understanding.
But one of the things AI is good at is looking at masses and masses of data. Most of the systems are trained on data sets well beyond any human capacity. Could you really look at a couple of million pictures of chairs to understand what constitutes a chair? Humans don’t learn this way. We are far more intuitive and able to extrapolate from the specific to the general.
When it comes to politics, having an AI provide a consensus opinion for a particular riding might sound like a good idea. By scanning Facebook posting, Twitter feed, conversations in public spaces, smartphone texts, and the rest of our electronic communication, it is possible an AI could come up with a better understanding of how a particular community truly feels about an issue.
But how would it then parse the dichotomy of opinions into a single point of view? Wouldn’t the result always be “well, I am of two minds on the issue…”?
This is an issue also faced by politicians. For the most part, I would suggest they resolve the issue by considering their own opinion, their own values, and the view held by the party. Sometimes the last matters most so politicians find themselves voting on a bill or acting in a fashion that is at odds with their own personal convictions.
But they are able to resolve these moral compromises. Or, at least, learn to live with the choices they made. Can a computer? Would an AI even have morals to compromise? Would it care?
If democracy or government is broken – and I am not saying it is – then the only way to fix it is by participating. Getting out and talking to friends and strangers. Discussing the important issues. Being involved. Letting our politicians know what you think and how you feel.
Perhaps with enough information, they might make better choices.
