I really wanted to join the AI Resistance Movement when I saw this beautifully designed masterpiece. After all, who wouldn’t want to be like some kind of Martin Luther, nailing those 95 theses on the church door?

Five hundred years ago, on Oct. 31, 1517, the small-town monk Martin Luther marched up to the castle church in Wittenberg and nailed his 95 Theses to the door, thus lighting the flame of the Reformation — the split between the Catholic and Protestant churches. Luther’s act is taught as one of the cornerstones of world history, and remains a lasting symbol of resistance five centuries later.

But that’s not actually what happened…. (source)

It’s a bold move, one that members of the AI Resistance Movement might want to emulate. But there’s no door. No Nails. Just a website. You’ll find a beautiful image at the website link via this BlueSky post.

The Problem

The problem isn’t that refusing Generative AI might be a move destined for failure. We’re not in the Star Trek universe, fighting against the Borg.

Instead, the people pushing this are the ones who are also funding indispensable technology tools, like smartphones and many other technologies many of us can’t put down long enough to join a protest movement, make a sign, or whatever.

But even if our own reluctance and greed wasn’t the obstacle, the problem might be simpler than that. The Gen AI resistance lacks a compelling argument. The claims it makes aren’t really supported by evidence.

NOTE: The authors of the paper linked at the website above have objected to my running this through a ChatGPT Custom GPT. It’s hard to take it back. So I can only offer my apology for not asking first. To make up for it, I’m going to remove the AI assessment that appeared below.

Here is the complaint from the authors:

Having complained about this myself (in spite of the futility), I am sympathetic to the complaint. You can read my response (and counter-response) online. This offers me future reflection opportunities.

Wait, What About ChatGPT’s Memory?

I admit that I hadn’t asked how long ChatGPT might store data submitted via a Custom GPT like Skeptical Thinker. So, I asked it, “How long?” Here is the response:

OpenAI does not use data submitted to ChatGPT for training future models. Any information you input is not retained beyond your session, and it is not used to improve the model. However, always be cautious when sharing sensitive or proprietary information in any AI system.

Of course, you can when using ChatGPT, take advantage of temporary option:

You can also manage its memory (Settings->Personalization). Again, none of this is any guarantee of safety. Probably the only real way is to run local AI on your machine since you can control internet access.