For over 50 years, computer scientists have suggested that computers will become human. They will evolve from being able to solve simple mental tasks to finally consciousness. In the early days, Hubert Drefyus was one of the first to dispute these ideas, showing that the process by which computer scientists attempted to write computer programs does not reflect how humans think at all.
Enter Watson, a computer at IBM that recently beat human contestants on the popular game show Jeopardy. As Dreyus explains in the NYTimes, even with the new approach for artificial intelligence that Watson exploits, we are still not talking about thinking at all. In the end, how humans answer questions depends on what matters to them. The statistical methods that Watson uses do not mirror caring or meaning. That is, computers simply cannot and do not care.
No comments:
Post a Comment