AI and the Language Barrier (3 Minute Read)

Currently, the way AI assistants generate'lifelike' responses are based off fed data. Machines do not think for themselves. Using deep learning, multiple layers within the program powering AI assistants are automatically tweaked and improved based off of the enormous quantities of data it has been fed. For example, AlphaGo analysed millions of games of Go, and played itself thousands of times to understand which strategies and moves it could use to successfully win a game. It's the same with iFlytek's trial in Anhui Provincial Hospital, AI receptionists will be fed data regarding symptoms and possible diagnoses, and rely on this knowledge to direct its patients, the human language is very different.