Greg
Staff member
- Joined
- Nov 29, 2023
- Messages
- 0
There’s been a lot of talk lately about how AI can support agriculture.
At On The Box, we already use AI for back-office tasks like quality control. Two of our advisors also run Juno, an AI research start-up built by people who know the industry inside out.
That’s the key. Good AI using expertise, not just information from the internet.
Lately, I’ve seen people using general-purpose tools like ChatGPT or Claude for livestock forecasting. If I’m honest, that worries me.
Forecasting livestock markets takes more than data scraping. Tools like ChatGPT don’t have access to high-quality, detailed reports like Rabobank’s industry updates. They rely on summaries or surface-level public content. That often leads to AI-generated content that’s shallow, incomplete, or flat-out wrong (otherwise known as AI slop)
Experts like Tim McRae (who works with us so I am biased) the EP3 guys Matt Dalgleish and Andrew Whitelaw, Angus Gidley-Baird , and others don’t just share data. They interpret it and help the rest of us make sense of what’s actually going on. That can not be replicated by generic AI without industry-specific models and training.
To show what I mean, I asked ChatGPT why Manly lost the 2007 NRL Grand Final. This is a sore point for me. I’m a self-proclaimed Manly tragic and yes, it still keeps me up at night.
The answer seemed reasonable at first, until I mentioned Melbourne was over the salary cap. Then it changed its mind. I tried again and said maybe experience was the key, and the answer changed again.
That’s the thing. AI can be steered to tell you what you want to hear.
So what does it all mean? AI could definitely have a role in ag and forecasting, but if we’re not careful it could do more harm than good. This is something we need to be cautious about, especially when recommending tools to others.
That’s just my take. What do you think?
At On The Box, we already use AI for back-office tasks like quality control. Two of our advisors also run Juno, an AI research start-up built by people who know the industry inside out.
That’s the key. Good AI using expertise, not just information from the internet.
Lately, I’ve seen people using general-purpose tools like ChatGPT or Claude for livestock forecasting. If I’m honest, that worries me.
Forecasting livestock markets takes more than data scraping. Tools like ChatGPT don’t have access to high-quality, detailed reports like Rabobank’s industry updates. They rely on summaries or surface-level public content. That often leads to AI-generated content that’s shallow, incomplete, or flat-out wrong (otherwise known as AI slop)
Experts like Tim McRae (who works with us so I am biased) the EP3 guys Matt Dalgleish and Andrew Whitelaw, Angus Gidley-Baird , and others don’t just share data. They interpret it and help the rest of us make sense of what’s actually going on. That can not be replicated by generic AI without industry-specific models and training.
To show what I mean, I asked ChatGPT why Manly lost the 2007 NRL Grand Final. This is a sore point for me. I’m a self-proclaimed Manly tragic and yes, it still keeps me up at night.
The answer seemed reasonable at first, until I mentioned Melbourne was over the salary cap. Then it changed its mind. I tried again and said maybe experience was the key, and the answer changed again.
That’s the thing. AI can be steered to tell you what you want to hear.
So what does it all mean? AI could definitely have a role in ag and forecasting, but if we’re not careful it could do more harm than good. This is something we need to be cautious about, especially when recommending tools to others.
That’s just my take. What do you think?