AI is making teaching history much harder (Patreon)
Content
As I write next week's newsletter, I'm on the fence about including this so I figured I'd start by sharing it here on Patreon. Essentially, it's me airing a frustration that I've encountered over the past few months.
Let me start by saying, I love my job. I may have the best job in the world. Researching food history and getting to share what I've learned with you each week is truly living the dream. But it's getting harder and I blame it on AI.
This past month, my facebook and instagram feeds have become inundated with accounts which discuss food history. At first, I imagined each account was run by a food history lover like myself. Some amazing nerd typing away to share with the internet, the latest historical factoid they'd learned. I was, I have realized, naive. After paying even a little attention to what these accounts are putting out, I've come to see they are entirely run by AI. The images are AI, the writing is AI, and worst of all, the research is done by AI. This wouldn't be a problem if AI was competent, but when it comes to history, AI is a mess.
I won't pretend to know how things like ChatGPT actually work but what I do know is they seem to trust anything that is told to them. They don't know how to fact check and, when teaching history, you have to fact check! Many of these AI accounts are sharing "facts" that are so quickly debunked that it's clear there is no person even trying to check to see if what they're posting is real.
History is hard. For hundreds of years, myths and mistakes have weedled their way into the history books so that it's often impossible to tell what is correct and what is not. I've made mistakes, the Smithsonian has made mistakes, everyone makes mistakes. I forgive these kinds of mistakes because, when a falsehood is repeated by a reputable source, it's only natural to believe it. But the mistakes in these AI posts are often so outlandish that it's hard to believe anyone could have signed off on hitting the post button. The issue is, when I look at the comments, it's clear that hundreds and thousands of people are believing these "facts". Some have become so popular in recent weeks, like the story of how George Crum invented french fries (there's so much wrong with this, most of which I covered in the episode on the history of potato chips), that people have been commenting on my YouTube video to incorrectly correct it.
My worry is this is just the beginning. ChatGPT has only been out for a couple of years and already the majority of images and stories that I find when I first start researching an episode are nothing but AI slop. What will it be like in another 2 years? Or 10 years? At what point will the real history, the well-researched, well-written, and fact checked history, be overwhelmed and drowned out by AI writing that is merely history adjacent? Close enough to be believed but not close enough to be true.
Maybe I'm being a bit of an alarmist. Perhaps someone will come along and tweak the algorithm scrubbing the vast amount of false information from the internet. But based on the popularity of these accounts which have spread to Instagram, TikTok, even YouTube, I don't have much hope. I, for one, will continue to do as I always have and research, to the best of my ability, filtering out the drivel and fact checking as much as I'm able. But it has become harder and it will continue to do so and I just hope that the falsehoods don't consume us all.
Thus endeth the rant.