LLM Trust Trap: Why Natural Language Bypasses Critical Thinking | Skin in the Game with Amit Shivpuja

There is a paradox emerging in the age of Generative AI.Business leaders are trained to be skeptical of raw data. They question spreadsheets. They interrogate dashboards. But the moment that same data is wrapped in a conversational sentence by an LLM? The skepticism vanishes.In the latest episode of Skin in the Game, I sat down with Amit Shivpuja (Director of Data Product & AI Enablement @ Walmart) to discuss why our biggest AI challenge isn't the technology—it's the misplaced trust that comes with a friendly chat interface.As Amit put it:"People tend to put a lot more trust in the system than they would normally do. They're a little skeptical if you give them a table of numbers, but if you write them a sentence saying the answer is X, they tend to trust it. It's on us to show them how we came up with the answer, not just give them a number."We also explored a fascinating case from his VR days where survey data (what people said) completely contradicted usage data (what people did). The lesson? Whether you are dealing with human feedback or AI output, validation is non-negotiable.If your organization is racing to adopt AI tools, this conversation is a vital reality check. We dig into:The "Data Sandwich" framework for building trustworthy systemsHow to build trust through radical transparencyWhy accountability must always remain with the human🔗 Links & Resources:linkedin.com/in/amitshivpujadatacompass.substack.com📺 Watch on YouTube: youtu.be/o6n5IP_Z2DYWhat makes YOU trust an AI output? Is it the authority of the tone, or the transparency of the source?. Drop your thoughts below. 👇#AI #DataGovernance #DigitalTransformation #Leadership #GenerativeAI #Walmart #DataIntegrity #StanfordLEAD #stanfordgsb #SkinInTheGamewithDebbieGoEnjoying the podcast? Leave us a quick review!Support the showFollow Us:🎧 Podcast website 💼 LinkedIn ▶️ YouTube