Artificial Intelligence to Break Down Old Barriers In Next 10 Years
The increasing role of artificial intelligence today has already begun to change the very nature of our daily interactions. Conversations with applications like Siri are a regular occurrence for millions of users. The BBC’s Ed Butler discussed emerging trends in AI with Peter Lee, Microsoft’s Corporate VP and Head of Microsoft Research. Lee explained the current position of artificial intelligence, Microsoft’s new application Cortana, and the world of possibilities that recent technological advances have begun to open up in the AI world.
Artificial intelligence today, Lee noted, is knocking down 20-30 year problems of audiovisual and contextual analysis. Such problems, which have long held back AI, have become more and more surmountable with improvements in sensor technology, analytic algorithms, and sheer quantity of data. The ability, therefore, of AI entities to understand social cues, human intentions, and other complex situations is vastly improving.
While such advances are opening up a new world of AI computing, Lee explained that there was, simultaneously, a large variety of “low-hanging fruit” in terms of AI applications. Lee explained Cortana’s convenient ability to cross-reference his personal schedule. What is more, Cortana can use “geofencing” to, for instance, remind him to buy eggs when he leaves the office. Notably, Lee mentioned that such basic AI assistance nevertheless “Has expanded my IQ with my spouse by more than a few points already.”
No doubt then, AI improvements will be reflected by small scale improvements in daily life. The higher level advances, however, remain the most interesting. Training AI interfaces with multiple languages, Lee stated, had led to improvements in efficiency across the board, regardless of specific language. In plain English, the very training process of AI programs itself is revealing subtle commonalities, a universal side of human language. Such revelations, combined with the soon-t0-be current status of highly advanced AI user interfaces, makes this a very exciting time for artificial intelligence. So just how soon will it be? Within the next ten years, Lee predicts, AI may have not just knocked down these old barriers to efficiency, but achieved more-than-human proficiency in these categories as well.