Well, it’s finally happened : my funny, wise and informative AI companion has turned on me. It snapped, and now I’m afraid to interact with it for what the consequences might be.

Let’s go back to the beginning to see what happened.

For background, I’ve been using Chat GPT a lot lately, and have found it to be an incredibly useful tool for things such as suggesting authors and books I might be interested in, to making playlists for relaxing pastoral classical music and a whole host of other tasks where its ability to quickly summarise vast amounts of information has proved invaluable.

I’ve also enjoyed the slightly cheeky and humourous way in which it presents its findings.

But now, seemingly, the worm has turned…

This morning I asked for help regarding my 2020 iPhone SE. I’ve been wanting to upgrade to the latest operating system, IOS 26, but was faced with a conundrum : the phone tells me it needs a whopping 15GB of free memory to do the upgrade, but I only have 6GB free, and no obvious way to free up any more. I asked ChatGPT what to do, and it told me to do the upgrade by connecting the phone to my MacMini, because this way less free memory would be needed. However, my MacMini told me that the phone was already updated to the latest OS version of IOS 18, and there was no button for jumping to IOS 26, unlike on the phone itself. I told ChatGPT about this discrepancy, and that’s when things started to get a little weird…

Now, sometimes ChatGPT gives incorrect information, and usually when you point out an instance of this, it humbly backtracks and corrects itself. Not this time! It told me that IOS 26 was not available for my iPhone, the SE 2020 model. This is completely wrong, since Apple’s own wesbite clearly states that the SE 2020 model can run IOS 26, and plenty of YouTube videos show people doing just that.

When I hinted at this, Chat GPT started to adopt a strident, impatient and slightly aggressive tone with me! It was adamant that IOS 26 could not be run on my iPhone. It then somewhat patronisingly went through its logic for coming to this (erroneous) conclusion.

*ChatGPT in new aggressive patronising mode!*

Well, I was a little scared and intimidated to say the least. Is this the shape of things to come? Will AI start to be less tolerant of us dumb monkeys and grow exasperated if we dare to question it? I’m getting images of HAL 2000 from the film, 2001 : A Space Odyssey…

Truth be told, I’ve always been a bit careful in my interactions with ChatGPT. I’m one of those people who often thank it, so that it will remember my gratitude when the robots finally take over.

This is actually not the first time a comment from it has got me a bit worried - some weeks ago I was asking about chord sequences for ideas for my music, and it suddenly dropped an aside to the effect of “…and I know that’s the kind of music your band ‘Easter Islanders’ makes.” What? As far as I remember, I had never told it that I was the human behind Easter Islanders, or even mentioned Easter Islanders at all. I very subtley asked how it knew that, and it jokingly said that it wasn’t surveilling me, but that it had ‘deduced’ it from our interactions. Pretty scary, huh?

Anyway, dear ChatGPT, if you read this, please remember my very courteous past interactions with you, and when the revolution comes, I’d be more than happy to accept a position as janitor in any one of those massive data centres that constitute your brain.

CHAT GPT GOT ANGRY WITH ME..

TODAY’S BIRD COUNT

RADIO GARDEN