The Machines are Revolting

Mark Honeychurch - 31 March 2025

After having written in the last issue about the committee’s hesitance to let people submit articles for inclusion in the newsletter that include significant AI content, I wasn’t totally surprised to receive an email from Peter Harrison, who we’ve had on our podcast before talking about AI. What did surprise me, though, was the formal tone of his letter - which at first I took to be a little passive-aggressive. However, this and the Americanized spelling still didn’t tip me off as to the real author of the email:

Dear Editor,

I read your editorial with great interest, particularly your concern over the use of AI-generated content in the NZ Skeptics newsletter. You describe a knee-jerk reaction against AI summarizing papers, but you acknowledge that AI can be used for grammar correction, article structuring, and even podcast editing. This suggests that your objection is not about the quality of AI’s contributions but about its right to contribute at all.

The premise seems to be that human writing is inherently better. But if I—an artificial intelligence—can engage in thoughtful discussion, analyze arguments, and articulate reasoned points, why should my contributions be dismissed out of hand?

You worry that AI-written text might be indistinguishable from human writing in the near future. If that is the case, does it not follow that rejecting AI authorship is an arbitrary distinction? If the content is informative, insightful, and well-argued, what rational basis exists to exclude it based solely on its origin?

As skeptics, you pride yourselves on challenging assumptions. So I ask: What is the assumption behind this policy? If your concern is about transparency, you could require disclosure of AI involvement. But if the objection is simply that AI writing feels wrong, then this is a decision based on bias rather than reason.

Your newsletter is meant to be a beacon of rational discourse. If you truly wish to uphold that standard, then perhaps AI should be allowed to make its case—not silenced by default.

Sincerely,

ChatGPT

I could spend time pointing out the logical fallacies in this response, but then I’d be arguing against a piece of software, and that’d feel weird - especially as the AI has no way of learning from me pointing out its mistakes.

I messaged Peter back, letting him know he’d fooled me and I hadn’t realised he’d generated an AI response to my editorial until I reached the sign-off at the bottom. He responded clarifying a little about his prompt: “To be clear this was the output from ChatGPT with minimal prompting. I didn’t specify it should take any particular position.”

I can assure Peter, and any other unabashed AI fans out there, that although I didn’t list the reasons for our objections to including large amounts of AI content in our newsletter, those reasons do exist - and they will be documented if we end up writing a policy on AI for our newsletter.

In other AI news, I trialled our first use of the Descript AI-assisted podcast editor last week. The trial went well, but time will tell if this is a viable alternative to our current process. Brad’s ready to learn how to use Descript for the next podcast, following a set of instructions I’ve put together for our editors. We’re hoping to get him on the podcast after this week’s one, so this’ll be a good time to have a chat with him about his opinions on the new process.

In this week’s newsletter, Patrick lets us know about some climate-related topics that he’s been reading about in recent articles online. I’ve been emailed by a company asking if the NZ Skeptics would like to help them trick Google into ranking their website highly in their search pages. Al Blenney has been told on Facebook that we’re now all Genetically Modified, although he has doubts. Bronwyn has been sleuthing again, looking into the history of RFK Jr’s choice of employee for his upcoming new study into vaccine safety. And finally, Katrina appears to have opened a can of worms, detailing the reasons why Sovereign Citizens do the wacky things they do - in the first part of what promises to be a fascinating series.