AI and Populism
This isn’t a post about how AI helps promote populist, far-right politics (although it does), but rather the similarities between the two with regards to the problems they pretend to address and the solutions they promote.
One of the persistent use cases for AI, and one that has actually seen some commercial uptake, is to help solve the problems caused by… AI. There are AI plagiarism detection tools, which can help institutions or markers detect possible use of AI, or help students ensure that their essay won’t be flagged for AI use. In job recruitment, AI helps people submit applications. The recruiter is now overwhelmed by applications, so they can subscribe to AI tools offered by sites to help filter out applicants, rank applications and reduce the number to a manageable load.
This is the perfect grift, making AI essential on both ends of the transaction. Individuals are told that they have to use AI to write their essay or complete their job applications, because everyone else is doing it. Organisations are told they have to use AI because the submissions they are getting are no longer reliable due to that pesky AI usage. At no point is “why don’t we just not use AI?” a seemingly valid response.
And so to populist politics. The approach is similar here in that the rhetoric is that only such politicians (Trump, Farage, etc) can solve the problems that they have largely created. By invoking anger and fear around immigration, they foster an environment that inevitably leads to violence and confrontation. And then, the soft liberals are all too weak to deal with it. As with Europe and Brexit, it was something that didn’t really feature much as an issue until Farage and co made it one and then suddenly the media play along and people are angry about something they barely registered before.
But the key here is that, as with AI, the solutions are fundamentally flawed. They blame immigrants, but when they come to power, simply removing illegal immigrants doesn’t solve any of the problems they said it would (because they were never the cause). So then they have to move on to other groups, each time saying “this is the one”: legal immigrants, anyone who isn’t white, left-wingers, trans people, the young, women. Pretty soon there is only about 20 people left on a tiny island shouting “it’s them!” The downward and increasingly violent spiral is guaranteed. So with AI, the systems can train on what the plagiarism or recruiting platforms are checking for and train on them. Just as the downward spiral is assured in populist politics, so the arms race of algorithms is the only outcome of the AI fed individual-organisation dynamic. It can never be won.
We do have some experience with defeating populism, although often that only comes after extreme damage has been done. But we saw in my neighbouring town of Caerphilly recently that viable opposition to the Reform rhetoric is viable, and as we’ve seen in Ireland and New York, the right-wing surge is not inevitable. What I wonder is if these victories can shed any light on the AI cycle? Are there tactics that were used, messaging that worked, values that people responded to, that can help prevent some of the AI slop tsunami? I don’t have the answer to those questions, just putting it out there as a possible route to think about it.
