Artificial intelligence is a domain that the vast majority of tech giants are researching. Facebook is also investing in researching artificial intelligence and it seems that they have results also. At the MIT Technology Review’s EmTech Conference held this week Yann LeCun, the director of Artificial Intelligence at Facebook showed the audience just how advanced the AI technology is on Facebook M. What is Facebook M you ask? Facebook M is the artificial intelligence assistant developed by Facebook. It is similar to Apple’s Siri and Microsoft’s Cortana but it is a lot more advanced.
Facebook started to develop M, the artificial intelligence assistant in 2013. Although in artificial intelligence research two years is not exactly a long period of time, Facebook already has great results to show off. Facebook’s M is able to answer questions in a way Siri or Cortana couldn’t. Facebook M is not only able to give you directions or solve simple tasks, Facebook is trying to make Facebook M more clever than a human being. And this is definitely not a simple task to complete. This is why at Facebook Artificial Intelligence Research experts are using humans to operate the AI powering Facebook M. Thanks to this Facebook’s M is able to complete complex, multi- steps tasks.
There is a problem with the human element in the program though. Because Facebook M’s AI is operated by humans, for now for now it is impossible to deliver the AI assistant to the over 1.5 billion users of Facebook. Researchers still need some time to figure out exactly what people expect from their AI assistant. Only before this will become clear will they be able to develop Facebook M to be able to give users exactly what they need. In order to have a good AI assistant in M users will also need to be willing to share a great deal of personal information with Facebook’s M. Given the cyber security right now this might be a little bit tricky. People are more concerned with what they share with or on their mobile devices. Or at least they should be.
Although there is still a long way ahead of Facebook’s Artificial Intelligence Research group, the result obtained so far are impressive. At the MIT conference Facebook’s Yann LeCun proved that M is already able to answer to questions asked by users. Facebook’s M is able to recognize images and answer questions based on them. It is also able to process images and make connections that AI assistants couldn’t before. Based on a sink for example, Facebook M was able to figure out that one picture was taken in a bathroom. Although it may not seem as a great step forward, I think it is. It may even be possible for AI assistants to become reliable helps for people that can’t see. If this becomes reality, it will be great. So far we just have to see how the tech giants research the domain of artificial intelligence and what use they will put artificial intelligence to. We just hope they will be able to use their research for the greater good.