Garbage In, Garbage Out
What We Can Learn About AI from a Farmer
Yesterday I found myself at a Climate Summit listening to a panel on sustainable food & agriculture innovation. When the topic turned to the role of AI in the space, the most thought-provoking insight came from someone who many might not have expected: a farmer.
Mark Shepard, a regenerative agriculture expert, was on a panel discussing innovative ways that farmers around the world are rethinking the meaning of agriculture. (For those interested in the space, I would HIGHLY recommend learning more about his work). One of Mark’s more prescient points was about how AI is actually preventing innovative agriculture practices from breaking through because of the way AI works — or in many ways, doesn’t work — as a model.
The truth is, isn’t just an agriculture issue — this affects every single person and industry that has make AI an integral part of the way they do business.
AI is the topic of the moment. Across industries and functions, leaders are exploring how to use it to automate tasks, generate content, and analyze data faster than ever before. It’s incredibly powerful, but it also opens up risk that many people don’t understand.
The risk I’m talking about isn’t some sci-fi fantasy that the AI “goes rogue” or replaces human jobs overnight. It’s something more fundamental and relevant to today: model collapse.
Model collapse is what happens when AI systems begin to reinforce flawed logic because they’re trained on outputs of other models instead of original, high-quality data. Basically, they use data that is created by other AI tools instead of from the source. It’s an echo chamber of automation that can look a lot like a game of “telephone” many of us played as children.
When bad data goes in, bad answers come out. But because they’re delivered by sophisticated models, we assume they’re valid. That’s a dangerous kind of confidence.
This is exactly why we at Zero Point Strategy emphasize starting from the zero point—that foundational moment of clarity where bias, outdated assumptions, and noise are stripped away. Before innovation, before growth, before transformation—there must be clarity. Because if your baseline data is wrong, your smart strategies will still lead you off course. AI doesn’t solve that problem; in fact, it can amplify it.
The lesson from Mark Shepard is this: even the most advanced tools can’t replace the need for understanding your ground truth. In agriculture or in business strategy, what you feed into the system determines what grows out of it.
If you're incorporating AI into your work, here’s what to remember:
KEY TAKEAWAYS
Interrogate the Inputs: Before trusting the output of any AI tool, ask where the data comes from—and who (or what) generated it.
Don't Automate Assumptions: AI can scale your thinking, but it can also scale your blind spots. Review decisions with fresh eyes.
Know Your Terrain: AI should serve your strategic context, not overwrite it. The best insights still start with human judgment.
Look for the Zero Point: Step back and strip away the noise. Find the moments where clarity leads, not complexity.
Treat AI as a Partner, Not a Proxy: It can inform your work, but it shouldn’t replace the deep, critical thinking that drives innovation forward.
