Philipp Schmitt’s “Blueprints Of Intelligence” is an incisive essay that examines how artificial intelligence (AI) has been portrayed over the years and the implications of these depictions for our understanding and conception of intelligence. The essay explores the history and evolution of AI, from its inception to its current state, with a particular emphasis on how AI and its components have been depicted visually.
Schmitt begins the narrative with a moving story about the brain autopsy of Albert Einstein. He contrasts the concept of physical brain examination with the approach of AI researchers, who seek to recreate intellect without dissecting a physical organ. This functions as a metaphorical foundation for the remainder of the essay, laying the groundwork for investigating how AI is visualized and how these visualizations contribute to our perception of intelligence.
The essay then explores how AI is frequently portrayed in popular culture, primarily through sci-fi motifs involving humanoid robots, circuit boards, and binary code with illustrations of brains. Despite their visual appeal, Despite their visual appeal, Despite their visual appeal, Schmitt contends that these images may mislead the public and impede productive discourse about artificial intelligence. He suggests that these images present a limited and potentially harmful perspective of artificial intelligence, disregarding the complexities of bias, ethical considerations, and the nature of intelligence itself.
Schmitt contrasts this with the visual representation of AI in academic contexts, where neural network diagrams and mathematical visualizations are frequently used. These diagrams, he argues, are the genuine “blueprints of intelligence.” These diagrams provide a functional representation of the underlying structure and mechanics of AI systems, as opposed to being aesthetically appealing or evoking emotion.
He traces the evolution of these academic depictions of AI throughout the essay, from hand-drawn neurons reminiscent of Santiago Ramón y Cajal’s neural brain drawings to creative representations such as Oliver Selfridge’s “Pandemonium” model to the highly abstract, flowchart-like diagrams of contemporary deep learning neural networks. This evolution demonstrates an increasing abstraction and departure from biomorphic connotations, moving towards a more mechanistic and modular view of intelligence.
Schmitt poses an intriguing query at the conclusion of his essay: Are neural network diagrams a better representation of AI than sci-fi tropes? While acknowledging that most individuals cannot comprehend or relate to these diagrams as readily as they could to a humanoid robot, he suggests that this may be their greatest strength. These diagrams provide insight into how AI creators conceptualize cognition and may facilitate more fruitful discussions about the risks and opportunities in AI research.
“Blueprints Of Intelligence” is a reflective examination of the visual representation of AI and its implications. Schmitt encourages the reader to view AI not as a single entity but as a complex, evolving field with varying interpretations of what intelligence is. Through visual representation, he promotes a more in-depth understanding and nuanced discussion of the nature, purpose, and potential of artificial intelligence.