A long time dream in the field of computer science is the development of an artificial intelligence, a program that would think and imagine and grow and learn like a human being. It always seems so close and yet so far.
Were such a program invented, my entire field would be out of a job, as the installation of an AI would allow people to talk to their computer directly. Programming? The AI can do anything you ask. System administration? The AI knows how to run the machine, prevent it from crashing, and can debug in real-time. Potentially without even stopping the running program. Data entry? A robot with a scanner would enter an encyclopedia's worth of data in the time it takes a human worker to drink a cup of coffee.
Still, AI research continues due to the obvious benefits. While it has yet to actually produce such a program, it has not produced nothing. AI research that failed to develop emergent behavior still had useful applications in expert systems (which require human will and minds, but do provide useful information in coming to a decision about something). Obstacle avoiding has improved, and robots are for sale now that many people are willing to treat as organic animals. (Unofficial reports claim that owners of a popular robotic vacuum actually clean for the vacuum so the vacuum won't "feel bad.")
Visual recognition is getting closer to useful. Machines still can't exactly "see," but they can associate identities with faces, which would have been outright impossible 30 years ago.
Work with trying to get machines to comprehend language has lead to both mechanical translation (which, while odd sounding, still gets the gist across most of the time) and word steming programs that lead to more natural feeling search engines.
Many experts claim that AI will be achieved incrementally, and no one breakthrough will actually achieve it, if it happens at all. So in a way, AI is both closer than ever (computers are smarter and can more easily help you), and farther away than ever (my job is secure for the forseeable future).
No comments:
Post a Comment