Class 1: Can Machines Think?
We begin with the foundational question that launched artificial intelligence as a scientific field: can machines think? Students explore the groundbreaking ideas of Alan Turing, whose famous Turing Test reframed intelligence as something that could be evaluated through conversation rather than consciousness. After examining the early optimism and setbacks of AI research, students conduct their own modern Turing-style experiment using large language models such as Chat or Claude. Rather than simply chatting, they design structured prompts to probe reasoning, memory, creativity, and consistency. The goal is to move beyond novelty and into analysis — understanding where machines simulate intelligence convincingly and where the illusion breaks down.
Class 3: Writing with AI
Large language models have transformed how writing is generated, revised, and distributed. Inspired by industry leaders such as Sam Altman, students use Chat or Claude to generate stories, persuasive essays, and scripted dialogue. The focus, however, is not passive generation — it is deliberate control. Students refine prompts with increasing precision, adding stylistic constraints, structural rules, and audience specifications to see how outputs evolve. Along the way, they analyze how predictive language systems construct meaning from statistical probability rather than lived experience, gaining practical fluency in guiding AI while maintaining critical judgment.
Class 5: Composing with Code
Music, once thought to be uniquely human, is now being generated by algorithms trained on vast libraries of sound. Drawing from research by François Pachet, students use Suno to compose original pieces across genres and moods. By modifying tempo, instrumentation, lyrical direction, and emotional tone, they observe how small prompt changes influence musical structure. Students compare AI-generated compositions with human works, examining repetition, variation, and emotional impact. The class explores whether creativity can be modeled mathematically — and where intuition and experience still distinguish human artistry.
Class 7: Coding with an AI Partner
AI is no longer just a subject of study — it is becoming a collaborator in software development. Guided by the educational work of Andrew Ng, students use Replit’s AI-assisted coding tools to generate and refine simple interactive programs. They experiment with building games or quizzes, then test and debug the output, comparing AI-generated solutions with rule-based logic they design themselves. This class emphasizes iteration, oversight, and evaluation, showing how human intention and machine suggestion interact in modern programming workflows.
Class 2: How Machines Learn
Modern AI systems do not follow rigid rules; they learn from data. This class explores the neural network revolution popularized by researchers such as Geoffrey Hinton, whose work laid the foundation for today’s deep learning models. Students examine how training data, pattern recognition, and layered computation enable machines to classify images, predict text, and make decisions. Through hands-on experiments with language models — and optional neural network simulations — students adjust datasets and observe how outputs shift in response. By seeing how easily bias or imbalance can distort results, they develop a deeper understanding of both the power and limitations of machine learning systems.
Class 4: Seeing Like a Machine
Artificial intelligence now interprets and creates images with remarkable realism. Building on breakthroughs in computer vision pioneered by Fei-Fei Li, this class explores how machines “see” through large-scale image training. Students use Higgsfield tools to generate hyper-realistic scenes, stylized art, and historical reconstructions from text prompts. They then participate in an AI detective lab, comparing authentic photographs with AI-generated images and identifying subtle artifacts in lighting, anatomy, and texture. This session blends creative experimentation with visual literacy, helping students understand how training data shapes machine perception.
Class 6: Moving Images & Intelligent Systems
Artificial intelligence increasingly operates beyond text and images, shaping dynamic media and decision-making systems. Influenced by the work of Demis Hassabis and DeepMind’s breakthroughs in reinforcement learning, this class explores how machines learn through reward-based optimization rather than explicit instructions. Students use Higgsfield’s video tools to create short AI-generated scenes and analyze motion coherence, realism, and narrative consistency. Through simulation-based examples, they observe how systems improve through feedback loops, gaining insight into how AI learns to act in complex environments.
Class 8: AI Safety, Risk & Responsibility
As artificial intelligence systems grow more powerful, questions about long-term safety and alignment become increasingly urgent. Drawing on the research of Roman Yampolskiy, this final class examines the risks, uncertainties, and ethical considerations surrounding advanced AI systems. Students analyze real-world applications such as recommendation engines, automated hiring systems, and autonomous decision-making tools. They discuss alignment, control, and unintended consequences, and present a capstone project combining multiple AI modalities. The course concludes by challenging students to consider not only what AI can do, but what it should do — and who is responsible for ensuring that outcome.
