Leadership Playbook | Episode - 5| Future of AI & How AI Really Thinks Explained by Tom Griffiths скачать в хорошем качестве

Leadership Playbook | Episode - 5| Future of AI & How AI Really Thinks Explained by Tom Griffiths

The future of artificial intelligence is transforming how humans learn, reason, and make decisions. In this episode of Leadership Playbook, cognitive scientist Tom Griffiths explains jagged intelligence, AI mistakes and failures, and how language models actually work. This conversation breaks down neural networks in simple terms and explores the difference between machine learning and human learning — including why AI needs more data than humans. You’ll also understand AI strengths and weaknesses, AI bias explained simply, and how the AI confirmation bias problem affects business decision-making. We discuss real human vs machine reasoning examples, major AI adoption challenges, and how organizations can use AI productivity tools without blindly trusting AI outputs. The episode also dives into the future of AI reasoning and the rise of neurosymbolic AI — combining logic, probability, and neural networks to build more reliable intelligent systems. If you want a clear, practical understanding of the future of artificial intelligence, this episode provides insights on AI failures, bias, reasoning, and how AI will shape business and society. Subscribe for more leadership, AI, and future technology conversations. #FutureOfAI #ArtificialIntelligence #LeadershipPlaybook #AIExplained #machinelearning 00:06 Introduction to Tom Griffiths and his career journey 01:01 The origin of AI and the "Laws of Thought" 02:47 The three mathematical foundations of AI: Logic, Neural Networks, and Probability 03:34 Mathematics as a lens for understanding the human mind 05:11 The importance of the historical and philosophical journey of AI 06:26 Who is the audience for "The Laws of Thought"? 07:51 How modern AI systems (Neural Networks) actually learn 09:51 Key differences between human intelligence and machine intelligence 11:37 Evolution and the constraints on human cognition 16:04 The danger of applying human models of intelligence to AI 17:17 Misplaced trust and the "Jaggedness" of AI performance 18:58 Examples of AI failures: Token probability and number representation 20:58 How organizations should approach building AI labs 22:03 The rise of Metacognition: Managing AI systems 24:54 Cognitive Biases: Confirmation bias and sycophancy in AI 26:37 Implicit societal biases and stereotypes in AI models 29:19 Correcting misconceptions about AI scaling and limits 30:54 Advice for leaders: Balancing scaling with experimentation 32:36 The cyclical nature of AI research: From psychology to computer science 34:42 Conclusion and final thoughts