For many people, a future that includes artificial intelligence (AI) sounds just great, because it means helpful robots, self-driving cars, and virtual assistants instantly providing personalized information and guidance on any number of matters. Others worry that the robots and cars and assistants will become so smart that they stop following human directives, and chart their own future, perhaps without us. Somewhere in between, most rational people are interested enough in AI to want to know more about it, if only to satisfy their own curiosity, rather than determine whether they should embrace or fear it.
AI is currently one of the most fascinating, as well as one of the most confusing, areas of development in the larger field of Big Data and analytics (BDA). AI stimulates the imagination with visions of automating many human-like functions that have been historically beyond the reach of technology. What makes AI confusing is the variety and complexity of its methods, compounded by the recent tendency of BDA solution providers to apply the AI label to their analytic applications. Adding to the confusion are the strongly-held opinions of many AI researchers that are beginning to surface regarding its development.
The mainstream media is not much help, as it typically addresses only the most sensational AI developments (e.g., Watson Wins Jeopardy!2 Elon Musk Fears the Singularity!3)—and in only the most simplistic fashion. Attempting to learn more, one quickly encounters a formidable language barrier. This AI language barrier rises up in the form of arcane terms like “Hierarchical Hidden Markov Models” and “Inverse Reinforcement Learning,” which are the real names of two popular AI methods. People who can easily understand such terms tend to be those who read and write mathematical formulas, and who also understand concepts like stochastic processes and back propagation. The rest of us are left scratching our heads and looking for more accessible explanations of how AI does what it does, its current capabilities and limitations, and realistic assessments of how AI may be able to help or hurt us.
Electronic Access - Site License Fulfilled By Publisher
Electronic Access - Global site License Fulfilled By Publisher