Noticeably, researchers at Google focus on applying NLP algorithms to various services, including Google Assistant and Google Translation.

The expert further forecast humans will be able to have decent conversations with machines in five years, yet the quality will not be as precise as people talking to each other.

"Interaction is much more conversational through mobile devices," he said, noting that the team is working hard to apply text-to-speech synthesis to improve the quality of the conversation.

The experienced search relevance specialist at Google has 15 patents related to search algorithms and other technologies.

Choi Hyun-jeong, a linguist expert who recently joined the team, is working on 35 languages to improve the human-machine interaction.

"Korean is difficult for corpus building," she said, noting that spoken Korean language and written Korean language are very different.

She said Korean is difficult to master as Koreans do not use subjects in sentences during conversation.

"Korean is very unique," Choi said. "It is virtually impossible for machines to completely understand Korean at present." (Yonhap)