The Economist, noting the fear that artificial intelligence might alter election outcomes, by generating fake videos of candidates and otherwise spreading disinformation, reports that AI actually doesn’t seem to shift votes. Studies suggest that vote changing caused by AI amounts only to a few hundredths of a percentage point. The reason is that “voters tend to view all messages about politics as spam, whether they encounter them through television advertising, robocalls, websites or social media.”
AI-generated propaganda can boost the amount of fake news, and fake news has political effects, per researchers and political operatives. But those effects are, first, to breed deeper distrust of everything, not just a particular side in the contest, and second, “to affirm (hyperpartisans’) views that are already strongly held.” Persuasion, as one researcher says, “is very difficult.”
AI, then, aside from streamlining “day to day tasks” of politics, mostly cements partisans in their existing animosities. It digs the trenches of partisan warfare deeper. But rallying one’s own side in intransigence and division has been an engrained political habit since the days before computers.
The studies of AI corroborate a view of current politics, cogently voiced in “The Politics Industry” by Michael Porter and Katherine Gehl (pp.23-4). The industry is a duopoly, of two parties tacitly splitting a market, in this case the electorate, and preserving their market segment, not truly competing for the “middle” but “prioritiz(ing) on each side the highly engaged …constituencies who most dependably vote or give money.”
Cultural conservative Patrick DeNeen also characterizes the two sides as “elite” birds of a feather, noting that both make little effort to say what is superior about their positions, instead “stress(ing) the vices of those they perceive themselves arrayed against.” In effect, the two sides need each other, as foils to rally their own partisans. In this sense, again, they behave as a duopoly, committed to preserving their franchise as their actual goal.
AI may well enhance the parties’ capacities to dissect an issue into “red” and “blue” sides. This almost ranks as a day-to-day task, keeping ongoing issues in their proper places. The two parties don’t need to define their brands. As this blogsite has noted, we are already inured to viewing political options as points on a single line defined by the two factions, ignoring any other points in three-dimensional reality. AI, like other marketing and communications tools, really serves to help the duopolists crowd out perspectives that could fuel other parties and movements.
Thus, as Porter and Gehl say, “the same two rivals are virtually guaranteed to remain in power no matter how poorly they serve the public interest. This would be a problem for customers in any industry – it’s a nightmare for a democracy.”
If third parties can’t make their mark from somewhere off the line between the poles, it may not all be the duopolists’ fault; we do have the vote. Still, if someone were to gain traction from “off the path,” there would be a deep and “bipartisan” reaction. Many Republicans’ as well as the Democrats’ rejection of Donald Trump seems, at least in part, to reflect duopolistic defense of their turf. And even as many on both sides denounce Trump, the duopoly still stands: Trump voters come mostly from one of the two sides, and the other exploits his bad qualities to rally itself.
America will not be served well until Americans serve themselves rather than accept two factions’ dominant mental constructs. If we see a higher set of values that we all share, the two sides devolve into options rather than identities, and their cultural/political outlooks into interpretations of America’s founding creed, rather than vindications of partisan “truths.” The Declaration’s creed gives us that bedrock of shared values, undergirding and, properly, overarching the duopoly. What the two options would implement, what they interpret for their competing brands, are merely different means to America’s existential ends.
AI today is just another tool in an old partisan game. We can change the game if we can view each argument and rallying cry as just a spin on our common creed, offered for politicians’ professional interests. AI bears examination as a public concern, but we can restore power to ourselves by putting propaganda, however it is generated, in its proper place.
Common Conviction 