Gerrymandering is one of the most insidious ways to influence our political process. By legally changing the way votes are collected and counted, results can be impacted, even fixed in advance for years. The solution could be an artificial intelligence system that attracts voting districts with an impartial hand.
Usually, districts that match electoral votes within a state are essentially drawn by hand, and partisan agents on both sides of the aisle have used the process to create distorted forms that exclude hostile voters and block their own. It is so effective that it has become commonplace, so much so that there is even a font consisting of letter-shaped districts.
What can be done? Automate it - at least in part, say Wendy Tam Cho and Bruce Cain in the latest issue of Science, which has a special section devoted to "democracy". Cho, who teaches at the University of Illinois at Urbana-Champaign, has pursued the reorganization of computational districts for years and just last year was an expert witness to an ACLU lawsuit that ended up overturning the destroyed districts of Ohio as unconstitutional.
In an essay explaining their work, he summarizes the approach like this:
The way forward is for people to work in partnership with machines to produce otherwise impossible results. To do this, we need to capitalize on the strengths and minimize the weaknesses of both artificial intelligence (AI) and human intelligence.
Machines improve and inform intelligent decision making by helping us navigate an unfathomable and complex information landscape. Left to fend for themselves, humans have proved unable to resist the temptation to trace precarious paths across that terrain.
There are actually an infinite number of ways to divide a state into a given number of shapes, so the AI agent must be triggered with criteria that limit those shapes. For example, perhaps a state doesn't want its districts to be larger than 150 square miles. But then they also have to take shape into account: you don't want one snake-like district crawling around the edges of others (as indeed often happens in cogwheel areas), or one being wrapped around another. And then there are the countless historical, geographic and demographic considerations.
In other words, while the logic of the design must be established by the people, it is the machines that must perform "the meticulous exploration of the astronomical number of ways in which a state can be divided."
Exactly how it would work will depend on the individual state, which will have its own rules and authority over how district maps are drawn. You understand the problem immediately: we have entered politics, another complex landscape through which human beings tend to "trace precarious paths".
Speaking to ProWellTech, Cho stressed that while automation has potential benefits for nearly all state processes, "transparency within that process is essential for developing and maintaining public trust and minimizing chances and perceptions. of prejudices ".
Some states have already adopted something similar, he pointed out: North Carolina ended up choosing at random from 1,000 computer-drawn maps. So there is definitely a precedent. But allowing widespread use means creating widespread trust, something that is in short supply today.
Mixing technology and politics has rarely proved easy, partly due to the invincible ignorance of our elected officials, and partly due to a justified distrust of systems that are difficult for the average citizen to understand and, if necessary, correct.
"The details of these models are complex and require a fair amount of knowledge in statistics, mathematics and computer science, but also an equally profound understanding of our functioning of our political institutions and the law," said Cho. "At the same time, while understanding all the details is daunting, I'm not sure if this level of understanding from the general public or politicians is necessary. The public generally believes in the science behind vaccines, DNA tests and flying planes without understanding the technical details. "
In fact, few people care if the wings will fall off their plane, but planes have proven their reliability over the course of a century or so. And the biggest vaccine challenge may be ahead of us.
"The company appears to have a huge confidence deficit at the moment, a fact that we have to work hard to reverse," admitted Cho. "Trust should and must be earned. We need to develop the processes that generate trust. "
But the point is: you don't need to be a statistician or machine learning expert to see that maps produced by these methods - peer-reviewed and ready to use, it must be said - are superior and infinitely more fair than many of those. whose borders are as crooked as the politicians who manipulated them.
The best way for the public to accept something is to see that it works and, like mail order voting, we already have some good points to showcase. The first, of course, is the North Carolina system, which shows that a fairground can be traced by a computer reliably, indeed so reliably that a thousand equally beautiful maps can easily be generated, so there is no. it's a selection problem.
Second, the Ohio case shows that maps can provide a fact-based contrast to gerrymandered ones, showing that their choices can only be explained by partisan meddling, not by chance or demographic constraints.
With AI it's usually wise to have a human in the loop, and doubly so with AI in politics. The roles of the automated system must be carefully forbidden, their limitations explained honestly, and their place within existing processes must be demonstrated to be the result of careful consideration rather than opportunity.
"The public needs to have a sense of reflection, contemplation and deliberation within the scientific community that produced these algorithms," said Cho.
These methods are unlikely to come into use anytime soon, but in the coming years, as maps are challenged and redesigned for other reasons, it could (and perhaps should) become a standard part of the process for an impartial system to take part in the process.