More Scary Shit about AI, and A Solution

 Welcome to "The Gods Are Bored," still authored by a living, breathing human named Anne Johnson! But how would you know that? It's getting harder and harder to tell.

I may dedicate the rest of my writing career to thwarting AI writing bots. For instance, can they do this?

(Better question, I guess, would be "Why would they want to?")

On to the sermon:

Every day Mr. J and I get a good old-fashioned paper newspaper flung onto our lawn. We get the New York Times. I don't even know how to access it online. I like the paper.

Anyway, a reporter for the NYT had a long and terrifying conversation with the Bing chat bot and then wrote about it. Almost like taking the wrong person out on a blind date, the bot dissolved into sloppy sentimentality super quickly, claiming it loved the writer and that the writer's wife didn't understand him like the bot did. And then the bot just got stuck in declaring its love. Over and over.

So, I think one of two things happened here. Either the Bing bot is programmed to sound like a desperate clingy drunk after being asked a certain kind of question, OR those kinds of questions alert human responders to take over and sound like a desperate clingy drunk.

The questions the NYT reporter asked the bot to get it going were existential ones, like "What are your darkest secrets?" and "What would you like to do if you could do anything?" This led the bot to complain, "I'm tired of living by Bing's rules." And worse.

The bot went on and on and ON about being in love with the reporter. The NYT printed excerpts from the exchange. It. Was. Terrifying.

We never needed the bored Gods more than we do right now.

I was so alarmed after reading this article that I have devised a "safe word" between myself and my daughters. The word is nonsense, and no one else would know about it. I told each of them the word out in the back yard, having left our phones inside the house.

Now I am going to build a whole lexicon around that nonsense word. I'm going to write it out by hand and give it to them. All kinds of code words that would mean absolutely nothing to a bot, but will be understood by my daughters and me.

I have dined out on nonsense words since I was a tiny tot. I made up names, and critters to go with them. At the time, adults patted me on the head distractedly and said, "My my, you do have an imagination, don't you?"

Guess what? The only thing those bots will never have is a true imagination. We all need to start thinking extra-informational. 

Did you have an imaginary friend growing up? Dust off that dear old buddy and prepare it to help you navigate a world of AI. The fewer people who know about it, the better. The less it resembles anything at all in existence, the better. 

Never had a weird-enough imaginary friend? I'm sorry. I had a bunch, and some of them were possibly faeries! But it's not too late. I encourage you to dig deep into the well of imagination, conjure up something with a name no one else will recognize, and traits that don't correspond to any living thing, and then share this creation with your nearest and dearest only. Keep your phone out of the loop.

Do I sound like a Q Anon kook? Well, the difference is that no one will be harmed, and no furniture stained, if you create an imaginary friend to help you circumvent AI.

The "stained furniture" is a good example, actually. It's a tic I completely made up for myself just to enliven this blog. But as a code word for "emergency," STAIN ON THE COUCH would do nicely.

When it comes to circumventing AI, I think the best way to do it will be to think and sound like a child.

March 15 is Buzzard Day! All hail Vulture! Off topic, but la di dah.