17/04/2026
🤖 ChatGPT: why I’m not worried about my job… but I am worried about your horses 🤖
I had a conversation with an agistment owner last week that stuck with me.
One of her agistees had asked ChatGPT for advice on her horse’s diet, and she wanted to double-check the answer.
The problem was, I didn’t have the full picture.
And I don’t guess.
If I’m going to give advice, I want the full diet, the hay, the pasture, the workload, the history.
So we went back and forth. I asked questions, filled in gaps, and eventually got to a point where I was comfortable giving an answer.
I answered the question because I’m a nice person who actually cares.
AI doesn’t. It can’t. It’s not a person.
What surprised me wasn’t the question, though.
It was the outcome.
Despite having the chance to get advice from a qualified human who will ask questions, look for gaps, and tailor things properly… she chose to follow the answer from something that can’t see the horse, can’t assess nuance, and didn’t ask a single clarifying question.
And that’s the bit that worries me.
Because AI only works with what it’s given.
To get anything close to useful advice, you’d need AT LEAST this input:
– Hay details (and ideally an analysis, but even then…)
– Pasture type, length, and current seasonal conditions
– Horse’s weight, age, condition score, workload, metabolic status
– Full supplement list (including doses)
– Clinical history (laminitis, ulcers, EMS, PPID, etc.)
And even then, there are gaps.
You might know the grass species in the hay.
But AI doesn’t understand how hay analysis shifts over seasons, years, and weather patterns.
After 10+ years of looking at forage, reviewing lab analyses, and watching how horses respond, you build a mental database.
So even without a lab analysis in front of me, I’m not guessing.
I’m making an informed estimate based on patterns I’ve seen play out over years.
AI might produce an answer in seconds.
But I can see the whole picture in seconds too… once I have the right information.
But it’s not just data.
I’m also picking things up from my gut.
From experience, from feel, from the horse in front of me, and from the owner and their attitude to that horse.
That combination of thought, feeling, and experience matters more than people realise.
AI can’t do that unless you explicitly feed it that context.
And most people can’t.
You can’t ask a question you don’t know exists.
That’s where experience comes in. It’s not just knowledge, it’s pattern recognition. It’s seeing the horse behind the numbers.
I’m not worried about my job.
Because feeding a horse properly isn’t just maths.
It’s judgement.
It’s context.
It’s experience built over years of watching what actually happens in real horses.
AI is a tool. It’s useful for education, for understanding concepts, even for asking better questions.
But it’s not a replacement for experience.
And your horse is the one that wears the difference.