Beginning with a Conclusion
October 17, 2024 Donald Trump Kamala Harris Nick Heer Culture Wars Post Truth Authenticity Generative AI Truth Slop
In an election already characterized by the culture war, half-truths, and manufactured controversies, it’s unsurprising that AI has a role to play. I nevertheless find it important to acknowledge that spammy and low-quality AI-generated content—what’s recently been called slop—is now actively shaping the way many people see the world.
An image shared by Trump on Truth Social
Donald Trump’s image of Kamala Harris speaking in front of a Soviet flag in Chicago is such an example. Blatantly faked, not even realistic-looking, it only supports his claim that the Democratic candidate is a communist—a tired stereotype, but one that still seems to find resonance.
There have, of course, always been fakes. But lately, it has become trivially easy to create such images that previously required at least a working knowledge of Photoshop. And with that inflation of slop, we’re seeing more and more manufactured images going viral and misleading people who either can’t tell they’re fake—or want to believe they’re true. More about that in a bit.
In my head, I keep turning over the phrase “manufacturing of doubt”, which I had heard on the excellent podcast The Last Archive some years ago. On the show, host Jill Lepore dives into the history of not just misinformation, but also looks at the people who stand to profit when there’s no more shared sense of reality. Doubt, I learned from the podcast, can be actively stoked—until we feel like we no longer know anything.
Hannah Arendt once wrote:
“Constant lying is not aimed at making the people believe a lie, but at ensuring that no one believes anything anymore. People that can no longer distinguish between truth and lies cannot distinguish between right and wrong… With such people, you can do whatever you want.”
The proliferation of low-quality slop for political purposes still has a limit: Not everyone will believe it; many images are easily debunked. But the pictures have a corrosive effect on the wider “information environment”: People start doubting that real things are fake, too.
Case in point: Another American election controversy, this time about the crowd size of a Kamala Harris rally shortly after she was elevated to the candidacy.
This time, Trump claimed that the image was AI-generated—in order to erode his opponent’s credibility. The catch: The image was real, but Trump skillfully tapped into the very doubt he had helped manufacture.
When Nick Heer wrote about this controversy, he showed it was based on a fringe claim from the darker corners of X.
This claim is so stupid because you do not need to rely on the photos released by the campaign. You can just go look up pictures for yourself, taken at a bunch of different angles by a bunch of different people with consistent lighting, logical crowds, and realistic hands.
But then he goes on and uses an explainer I find immensely useful:
I know none of the people promoting this theory are interested in facts. They began with a conclusion and are creating a story to fit, in spite of evidence to the contrary.
Beginning with a conclusion is key to what we’re so often seeing in the culture wars: Preexisting beliefs and grievances drive the fact finding instead of facts informing opinions people hold.
I think a lot of that has to do with the mechanics of the internet, with the algorithmic echo chambers that push people towards more content that keeps them upset and coming back for more—generating ad revenue in the process. But it’s also a human instinct to want to confirm one’s pre-held beliefs—and it’s never been easier to find you exact opinion replicated in some corner of the web.
Update (23 Jan 2025): In an article for The Atlantic, Charlie Warzel and Mike Caulfield write about this very mechanic, calling the internet a “justification machine”.
The justification machine,(…) has made the process of erasing cognitive dissonance far more efficient. Our current, fractured media ecosystem works far faster and with less friction than past iterations, providing on-demand evidence for consumers that is more tailored than even the most frenzied cable news broadcasts can offer.