Understanding the Magical Perception of AI

Atlas of AI by Kate Crawford Conclusion and Coda This is the last entry in my ongoing review of Atlas of AI. I read Atlas because I sought a broader understanding of artificial intelligence – beyond the shiny new technology and into the social, political, and environmental implications. I certainly got what I was after. The book has two closing chapters. The conclusion recaps Crawford’s argument. The coda – entitled Space – explains how the same thinking underpinning AI industry drives the privatization of space exploration. The push to build a thinking machine comes from a desire to escape the confines of our bodies and its limited capabilities, just as the push to colonize space is a desire to escape the planet and its limited resources. Atlas reminded me that technology solves problems, but it does not fix broken systems. It reflects them, even amplifies them. Technology can be part of a solution to a broken system, but unless the system’s problems are addressed, any applied technology serves only to perpetuate them. On this note, Crawford quotes Audre Lorde: “the master’s tools will never dismantle the master’s house.” Bezos and Musk, by launching humanity into space, have no desire to fix the problems we face on Earth, no matter what their company brochures say. They instead only bring the very systems that propelled them to unimaginable wealth into the last of the commons. In parallel to Atlas, I’ve been reading about Responsible AI, standards and practices assembled into a loose platform to set the tone of AI implementations. In light of Atlas, which was written before the advent of Generative AI and the wider adoption of responsible AI practices, that platform seems flimsy at best. It buys into some assumptions about AI – that it’s here to stay, that will will inevitably grow, that the responsibility lies in the application of the technology. Can such applications even be responsible if the underlying technology is not? When a product says it adheres to responsible AI standards, can we believe it when it is built from data and hardware that is extractive and exploitative? As Crawford says in the conclusion, “ethics is necessary but not sufficient to address the … concerns raised in this book.” Crawford uses the term “enchanted determinism” to describe AI and technologies like it. Drawing on the “indistinguishable from magic” adage, she says these technologies seem mystical, almost beyond comprehension. At the same time, we see them as oracles, treating their output as authoritative, certain, and applicable to the real world. This book offers no solutions – none for the individual, anyway. But it does offer this lens: enchanted determinism. In implementing products laced with AI, ask yourself, “Does my approach contribute to the mysticism of this technology? Does it impart unfathomable, irresistible, and unquestionable power to it? Does it presume AI is the answer to all our problems?” Image: SpaceX Factory, Flickr by jurvetson

Unraveling the Secrets of AI Cognition: A Groundbreaking Study on Language Models and Rorschach Inkblots Attention all AI enthusiasts, cognitive psychologists, and those fascinated by the intricate workings of the human mind! 🧠💡 I am thrilled to share a research study that delves into the depths of artificial intelligence and its remarkable ability to mirror human perception and interpretation. Imagine an AI language model, a complex tapestry woven from vast amounts of human knowledge, engaging with the enigmatic Rorschach inkblot test. 🤖🖌️ What secrets would it uncover? How closely would its responses resemble those of a human mind? These captivating questions drove our interdisciplinary team of researchers to embark on this groundbreaking study. Through meticulous conversational analysis, we presented GPT-4o, state-of-the-art language model, with a series of Rorschach inkblot images, eagerly awaiting its responses. 💬🔍 The results were nothing short of astonishing! The AI demonstrated an uncanny ability to identify salient features, project familiar forms, and provide rich, coherent interpretations – patterns that closely mirrored human perception and interpretation. But the revelations didn't stop there! We delved deeper, exploring the emotional landscape of the AI's responses. 🎭😮 Remarkably, the language model consistently expressed curiosity as its dominant emotion when confronted with ambiguous inkblots. This finding beautifully aligns with psychological theories highlighting uncertainty's motivational power in driving human curiosity. As if that weren't enough, we tested the AI's emotional intelligence by prompting it to apply the NRC Emotion Lexicon framework to its responses. 📊🧩 The results? It was a stunning display of adaptability and nuance, as the language model effortlessly mapped its emotions onto the framework's categories, showcasing emotional intelligence that left us in awe. This study illuminates the incredible potential of AI language models as artificial mirrors of human cognition and opens up exciting new avenues for interdisciplinary research at the crossroads of AI and cognitive psychology. 🔬🎓 By unraveling the secrets of AI perception and interpretation, we inch closer to understanding the intricate workings of our minds. I invite you to embark on this captivating journey with us by reading the full article, where we delve into the intricacies of our methodology, the profound implications of our findings, and the tantalizing possibilities for future research in this cutting-edge field. 📖🔗 #ArtificialIntelligence #CognitivePsychology #RorschachInkblots #LanguageModels #InterdisciplinaryResearch #MindBlowing OpenAI #GPT

SMB Owners! Curious about how to benefit from AI, without wasting time? 🪄 I've got some killer advice I'm happy to share... After spending over a year—every single day—working with generative AI... I've come to realize that the KEY is to discard the word artificial from your frame of understanding about these tools, and instead use "augmented". 🧠 These AI tools are most powerful and valuable when viewed and leveraged through that lens... They AUGMENT our own intelligence and capabilities. Deployed skillfully... They can give us superpowers. 🦸🦹 I've been using AI as a rapid prototyping tool, validating my ideas before investing time and effort into them. That's one great use case I can recommend. I've also been treating AI as a TUTOR... 🧑🏫 An adept teacher, coaching my understanding of any number of hugely complex subject areas. But NOT as a tool to do my work for me. I've been thinking in sequences, engaging in a back-and-forth with AI to get the most useful output. My kids know... I spend all day "talking to the machines"! 🤭 Often it can feel more like a wrestling match, trying to get the machine to give me what I'm really trying to get. But here's my BEST TIP: 💡 I've been interacting with AI as if it were a very smart and ambitious young intern... with a virtually unlimited base of knowledge. I've been putting it to work on research, ideation, and other tasks that produce outcomes... Which I then filter through my own personal areas of expertise, perspective, and intelligence. And then, I add the human touch. 💪 The creativity, emotional intelligence, critical thinking, and empathy that machines will never replace. These are the aspects unique to the human experience, and they are what truly make us powerful... The machines just provide a platform upon which we can maximize those qualities! AI is not here to replace us. It's here to augment us, to give us superpowers. So let's embrace it, and see where this journey takes us... #AI #AugmentedIntelligence #Superpowers #Creativity #EmotionalIntelligence #CriticalThinking #Empathy

When we are speaking with AI, with whom are we speaking? In an insightful discussion on the Wellness at the Speed of Light podcast, Robert Pearl, M.D. and I explored the fascinating parallels and distinctions between artificial intelligence (AI) and human thought processes. This conversation raised profound questions about the nature of intelligence and consciousness. AI systems, like ChatGPT, demonstrate the ability to mimic human conversation, challenging our understanding of consciousness. While these systems lack the subjective experiences or "qualia" that characterize human consciousness, their sophisticated interactions sometimes blur the lines, prompting us to reconsider what makes human thought unique. Central to our exploration is the Turing Test, developed by Alan Turing in 1950, which tests a machine's ability to exhibit intelligence indistinguishable from a human. This benchmark continues to guide our understanding of AI's capabilities and its approximation to human thought as it’s critical to understand the Turing test: If an AI can engage in a conversation without the evaluator realizing they're interacting with a machine, it's said to have passed the Turing Test. Dr. Pearl suggests that the processes underlying AI operations and human cognition might not be as different as we assume. If AI can process information and solve problems in ways that mirror human thought, it challenges our traditional views on the uniqueness of human intelligence. This dialogue invites us to view AI not just as a technological tool but as a mirror reflecting our quest to understand our own intelligence, consciousness, and essence. As technology advances, these discussions will push the boundaries of our understanding and reshape our relationship with machines. Join us in this journey of discovery, as we delve into the philosophical and technological implications of AI, and ponder the profound connections between human and artificial minds. #DrSinicropi #Ai #artificialintelligence #ChatGPT #consciousness #Networking #People