AI in the Real World! How AI Predicts Human Behavior | Galvin Widjaja

In this episode of An Hour of Innovation, host Vit Lyoshin sits down with Galvin Widjaja, founder and CEO of Lauretta.io, to explore how AI in the real world is moving beyond counting people to actually predicting human behavior.They discuss how computer vision AI is reshaping malls, airports, and casinos, why context is everything when it comes to intent detection, and how balancing privacy and ethics is critical for the future of AI surveillance.Galvin shares how Lauretta.io’s technology can detect suspicious behavior, from identifying someone planting a bomb without ever witnessing the act, to uncovering hidden demand in shopping malls, to supporting casino security systems that resemble something straight out of Ocean’s Eleven. He also explains why Loretta.io is designed to “forget who you are” by default, and why that ethical choice matters as predictive AI expands into the physical world.Galvin is the founder and CEO of Lauretta.io, an AI company that focuses on understanding human experiences in physical spaces through computer vision and non-biometric tracking. His background spans finance, consulting, and data optimization, which led him to spot early opportunities in AI surveillance and intent-based analytics. With a deep focus on AI ethics and privacy, Galvin is helping shape how predictive AI will be adopted in industries from retail to security. He also shares his advice for AI founders, thin vs thick AI, and why most AI startups will fail if they don’t choose wisely.Support This PodcastTo support our work, please check out our sponsors and get discounts: https://www.anhourofinnovation.com/sponsors/Takeaways* AI in the real world is shifting from counting people to predicting human intent.* Context matters: the same action means different things in different places.* The AI technology can infer behavior, like spotting someone planting a bomb, even if the act isn’t seen.* Retail use cases: understanding shopping behavior, hidden demand, and improving mall analytics.* Security use cases: airports, public spaces, and casinos with thousands of surveillance cameras.* AI acts as an assistant to human security guards, extending their attention span.* Lauretta.io is built to forget identities by design, prioritizing privacy and ethics.* Consent is key: identity is only linked when explicitly given or under legal security boundaries.* Embodied AI is coming, think robot dogs in homes that track movement and behavior.* Future AI will make intent detection more precise, but it raises ethical questions about preventive security.* Galvin stresses the importance of responsible AI design to avoid privacy abuses.* Galvin introduces the Thin vs Thick AI concepts: startups built on single models vs layered systems.* Most AI startups will fail if they rely too heavily on one public model.* Successful AI companies must focus on fundamental technology and long-term differentiation.Timestamps00:00 Introduction01:04 Galvin’s Journey into AI06:52 Ads08:06 AI in the Real World: Understanding Human Experience14:23 Lauretta.io Technology: Computer Vision & Non-Biometric Tracking18:06 Context Matters: How AI Predicts Human Intent22:53 AI for Retail vs Security: Business Insights & Threat Detection27:32 Casino Surveillance: The Ocean’s Eleven of AI33:11 Robot Dogs, Embodied AI & The Future of Surveillance39:06 Ethics and Privacy: Why Lauretta.io Forgets Who You Are43:25 The Future of Predictive AI & Security48:00 Advice for AI Founders & InnovatorsConnect with Galvin* Website: https://lauretta.io/ * LinkedIn: https://www.linkedin.com/in/galvinw/ * X: https://x.com/galvinw Connect with Vit* Website: https://vitlyoshin.com/contact/ * LinkedIn: https://www.linkedin.com/in/vit-lyoshin/ * X: https://x.com/vitlyoshin Vit’s Projects* Podcast: https://www.anhourofinnovation.com/ * AI Assistant to build apps: https://appforgelab.com/