Slop new world

By Ben Olson
Reader Staff

For the first time in the Reader’s history, we ran the candidate questionnaire responses that we publish before every election through an AI-detection software called Pangram. We were surprised and alarmed by the results, which led us to establish an AI policy to handle situations like this in the future.

Though five of the eight candidates submitted “100% human-written” responses to our questionnaire, Republican candidates Scott Herndon, Jane Sauter and Cornel Rasor all turned in responses that Pangram identified as “100% AI-written.”

This, in a word, was astounding. It was so alarming that we felt the need to put a disclaimer at the end of each candidate’s answers that were published in the Oct. 22 edition.

To clarify, when something is “100% AI-written,” it means that someone typed a prompt into an AI chatbot like Claude, ChatGPT or Gemini and then copy and pasted what the bot spit back verbatim. Pangram doesn’t just deliver review results of “100%” — it’ll show where and how a piece of writing was likely or certainly human-written, AI-written or AI-assisted, then break that down into various percentages and likelihoods. That didn’t happen here. These responses were 100% generated by AI.

It’s the most blatant case of phoning in an election that I’ve ever seen, especially when you take into account that Herndon, Jane Sauter and Rasor all declined to participate in the Oct. 20 election forum hosted by local media.

We can have discussions all day long about the ethics of using AI for writing, but I have to draw a firm line in the sand when we see evidence that politicians are using it to craft their own opinions on policies and issues that will end up affecting constituents if they get elected. In my view, it’s not just tacky or lazy, it’s a sign of contempt for voters. 

Here are a few hypotheticals: Would you be OK with your spouse generating wedding vows that are 100% AI-written? Would you accept a newspaper publishing 100% AI-written news stories that weren’t even reviewed by a reporter or editor?

Why, then, is it acceptable that politicians use it to such a degree?

In a state like Idaho, where lawmakers often lament the “out-of-state influence” on other legislators, nobody seems to be kicking up much of a fuss about three candidates who couldn’t even be bothered to formulate their own opinions.

It’s this, perhaps more than anything, that points to a broken system that’s getting worse and worse.

As a result, the Reader has developed a policy on AI submissions. Since the founding of this newspaper in 2004, we have always thought of our readers as smart, savvy people who are interested in authenticity, not rage bait or AI slop.

Therefore, here is the Reader’s AI policy, which will also be posted on our website:

1. The Reader will not publish any AI writing on our opinion, perspectives and letters to the editor pages. If we suspect AI was used to generate a submission, we’ll run it through a detector and, if our suspicions are found correct, we’ll return the submission for the “author” to rewrite it in their own words and resubmit. Otherwise, we will deny the piece.

2. Articles authored by Reader staff members have always been and will always be 100% human-written. We do not use AI to write anything, nor will we ever unless the topic of the article is specifically about AI writing and we are using examples. If AI is ever used for this purpose, it will be disclaimed at the beginning of the text.

3. We will not consider AI generated art for our cover page. If AI is detected, we will reject the submission.

4. While the majority of our AI detection will be focused on opinion articles and letters to the editor, we will also check other submissions if they are suspect and may reject them as well.

We believe in the power of the human mind. We believe that our readers deserve art, writing and thoughts generated by human beings, not bots. We believe in putting our heels down and holding the line. It’s that important.

Sadly, this is not just a problem here in North Idaho. As AI is becoming more accessible, more and more “creators” are using it without disclaimers, passing off slop as their own work. 

We have no idea how many articles we’ve already published that were partially or wholly AI written and, honestly, we’re a little afraid to find out. But that all stops now. 

The saddest part of this whole ordeal isn’t that AI is being used by politicians and opinion writers to craft articles that thousands of people read and remember. No, the saddest part for me is that no one seems to care.

Expect more out of those who represent you. When you stop caring, so do they, and it shows.