I liked seeing this report (more accurately, a reprinted press release) about Thinking with KIBO: Introducing Artificial Intelligence (AI) in Early Grades, released by KinderLab Robotics. I'm less enthusiastic, though, on seeing that two of the four learning objectives are factually inaccurate. One of them says "AI systems (and robots!) make decisions based on input and rules," but modern AI is manifestly not rules based, and it's misleading to say that it is. Second, it states, "AI systems don't really 'think' the way we do at all," but again, the neural networks used by modern AI are in many ways similar to human neural networks, and it is not at all clear that they represent two distinct ways of thinking, as opposed to variations on the same general theme.
Today: 2 Total: 1155 [Direct link] [Share]
