Most of us hear about Artificial Intelligence every day—but how does it actually work under the hood? In this episode of The Science Dilemma, Allan CP sits down with Dr. George Montañez, associate professor of computer science and director of the Amistad Lab, to break down what AI really is (and isn’t).

Dr. Montañez explains large language models with easy-to-grasp analogies (like rolling dice with words), reveals why AI often looks “smart” but doesn’t truly understand, and warns about the real danger—our misplaced trust in the technology.

You’ll also hear about:

• Why AI can mimic reasoning without ever understanding
• What Dr. Montañez calls AI Idolatry—and why it’s dangerous
• Practical guidelines for using AI responsibly (The Berean Principle)
• How diffusion models power image generators
• Opportunities for entrepreneurs to build with AI as tools, not replacements

This episode will sharpen how you think about AI, equip your family with clarity, and spark important conversations about technology, responsibility, and faith.

• Watch Dr. Montañez’s YouTube channel Theos Theory:
• 🎁 FREE Member PDF Here: https://sciencedilemma.mysamcart.com/free-member-packet
• Join our Podcast Community here: https://www.thesciencedilemma.com/podcast
• Purchase The Science Dilemma: Origin of Life here: https://www.thesciencedilemma.com

00:00 – Introduction: Why we’re the danger with AI

01:30 – Meet Dr. George Montañez

03:00 – AI explained with dice and tokens

05:30 – Why AI struggles with reasoning

08:00 – AI Idolatry: Misplaced trust in technology

10:40 – The Berean Principle: How to use AI responsibly

13:00 – Novelty vs. correctness: AI’s built-in trade-off

16:30 – Beyond LLMs: Image models, diffusion, and self-driving cars

20:30 – Opportunities to build with AI

23:30 – Stewardship, entrepreneurship, and serving others

24:10 – Closing thoughts & resources

AI explained simply, how does AI work, George Montañez, large language models, diffusion models, AI idolatry, Christian perspective on AI, Science Dilemma podcast