There's no doubt that Microsoft is betting big on the power of artificial intelligence (AI) for the future. As part of that bet, the company has been pushing forward with research that not only advances the capabilities of AI, but also explores future applications. In a new segment, BBC's Newsnight has gone inside Microsoft's AI research labs, giving us an interesting look at where the future of AI may take us. Check it out below:
As we propel into the future, Microsoft sees AI blurring the lines in a way that interacting with a computer feels as natural as interacting with another person. We've already seen some early attempts at this with the company's various chatbot efforts. Microsoft's Chinese chatbot, Xiaoice, for example, averages 23 turns per conversation, Dan Driscoll, who works on Microsoft's bot frameworks, says.
AI image recognition and natural language processing are also approaching levels that rival humans, largely thanks to advancements we've only started seeing in recent years, Microsoft explains. Recently, for example, a Microsoft research team was able to train AI to beat Ms. Pac-Man with a higher score than any human had previously obtained. While that may sound somewhat silly, it's part of important research that could eventually impact many different aspects of our everyday lives into the future. One example of the AI's potential covered by the BBC is in farming, where it could help reduce costs and resource consumption while increasing yield.
As touched on in the segment, there are legitimate concerns that AI could be leveraged by bad actors in the future. But, as Microsoft Research's Eric Horvitz explains, AI is also probably the best detector and counter to potential misuse.