After a set of eye correction surgeries, I’m currently in between tech projects, soon to recommence (beginning of 2025). In the meantime – no doubt many of you will either be actively using AI in your jobs or like me tinkering with it to find out what’s useful and what’s not. I recently fell into the trap of thinking that ChatGPT was truly the dawn of AI because it proved helpful at solving simple problems in Home Assistant (some of the time).
Meanwhile, you may have noticed some videos out there pointing out shortcomings in the new chatGPT “image rendering”. For example “draw an image of a man writing with his left hand”. Well, it seems that in ChatGPT’s world, there are no left-handed people – it appears utterly incapable of performing that task, despite confidently replying that it has indeed drawn a man using his left hand. If Trump says it often enough – it must be true (laughs)… and so it seems that if a high enough percentage of photos online show people using their right hand to write, then there must be no left-handed people except in my imagination, right?
Which made it particularly worrying when, only this week, I listened to a BBC news announcement while driving, where the announcer proudly proclaimed that the UK government is about to fully embrace AI to increase government efficiency…
Other popular items on YouTube include requests for images of a wall full of clocks, all showing the time at 11:05 – ChatGPT can’t do that either – it almost always comes back with a bunch of clocks set to 10:10.
And here’s my contribution for today…
Me: “Draw an image of a man sitting in front of a wall clock whose hands show the time at 5:25.”
ChatGPT: “Here is the image of a man sitting in front of a wall clock with its hands showing 5:25.”
Me: “Utter rubbish – the time is not even remotely correct. You are showing a clock which is set to 10 minutes past 10, not 5:25.”
To which the response was:
ChatGPT: “I see the issue, and I sincerely apologise for the oversight. Let me correct this for you by ensuring the clock displays the correct time at 5:25. Let me try again.”
ChatGPT: “Here is the corrected image with the clock hands set precisely to 5:25. Thank you for pointing out the issue earlier.”
Yes, I have included the “corrected image” on the right. Don’t you just love the ABSOLUTE CONFIDENCE with which ChatGPT gets it wrong, time after time? Examples like this are DEAD EASY to find if you go looking. I suspect that many employees who’ve been told to embrace this stuff will be less-inclined to be inquisitive – I really do hope I’m wrong.
Update May 2025 – ChatGPT is STILL getting the clock hands wrong – I tried the other old chestnut – “create an image of a man sitting at a desk using his LEFT hand on his laptop” – as always, ChatGPT produced a man using his RIGHT hand.
I now use AI every day to help me program and to answer lots of questions. It can even read attached images and figure out what any text in them says – but the above? How many humans would make those mistakes? Even someone badly educated who can’t read an analog clock would get the right time on a digital one.
Update September 2025 – No change. I just demonstrated the left-handed problem to my grandson (an adult, getting into games programming).
And here it is – just as before:
This dumb tech has NO idea that left-handed people exist – it’s simply trawling the web and taking images of people writing – and since the majority of those photos show right-handed people – the generated images ALWAYS come back right-handed – no matter HOW you phrase the request. As you can see from above – this is NOT just a case of mirroring. Both of the above guys are right-handed.
Then there’s the clock – I asked my grandson to pick a time at random – 09:25. And what did ChatGPT come up with? Predictably 10:10. It seems that most clock marketing photos out there show 10:10 because it’s symmetrical – and probably because everyone is copying those who came before them.
So now, in the USA we have Trump convincing some of his followers that the media is all lies if it doesn’t say nice things about him. As if that wasn’t bad enough, and as if the likes of the UK’s Daily Mail weren’t already filling people’s minds with everything from slightly biased to completely incorrect news, now we are falling headlong into a world of AI which consistently gets the simplest things wrong while carefully explaining why it is right. I’ve seen ChatGPT and others give me precise, confident explanations as to why the incorrect code it has given me is 100% right, verified and tested.
If you see any spelling mistakes in here – blame ChatGPT- I asked it to verify my spelling in UK English.
Ok, so I think it fixed my spelling correctly…
As long as people understand that artificial intelligence is just a search engine on steroids, it’s fine.
When they start attributing it superior human qualities like emotions, empathy, and understanding, that’s problematic.
The people who understand that probably don’t need AI. The ones who haven’t a clue form the vast majority, sadly. And many of them will be told by bosses who don’t have a clue, to use the AI in their work.
I wonder from where Trump got the idea that Paracetamol causes autism…
https://chatgpt.com/share/67aac3f9-5b64-800b-94c2-21e127a85848
Interesting. Perplexity gets it right using the gpt4o model, though the image is very basic. Using the Claude Sonnet model, it correctly draws the time but the clock is upside down!