Google Develops Artificial Intelligence With Human-Level Speech Capability

First Posted: May 31, 2016 10:10 AM EDT
Close

Google's artificial intelligence is reported to be developed under the supervision of Ray Kurzweil, a computer scientist and futurist. According to reports, he was hired by the company in 2012 to work on the natural language recognition.

Kurzweil, the man reported to be behind the Google artificial intelligence, showed in the the recent Singularity conference through a telepresence robot. In the same event, he announced that together with his team, he has been working on chatbox. He also revealed that the chatbox will be released later this year.

The announcement was brought up when his opinion was asked regarding the possibility of people having a meaningful conversation with an artificial intelligence, one that is not easily distinguished from a regular human conversation. Kurzeweil admitted that it is important to what he is doing at Google, and that his team is working on chatbots. He also added that they are expecting to release a few chatbots that people can talk to later this year.

Based on previous reports, Kurzweil is also known for his projections that pay attention on how technology and science can progress and modify the world. He is a called "prophet" of the Singularity that deals with the period when artificial intelligence will exceed humans in terms of ability and intellectual capacity. Of his 147 projections since the 90's, he has an 86 percent accuracy rate.

According to The Guardian, a chatbot named Danielle is based on a character from an unpublished book called Danielle, while its speech is patterned on the character's dialog. Kurzweil cited Danielle as a model of how people may produce their own unique chatbot by providing it a vast sample of writing, allowing the bot to adopt personality, style and ideas.  

Although he believes that his chatbots will allow for worthwhile conversation, he insists that it will take until 2029 for them to have the human-level language abilities. Kukrzweil also claims that, right now, artificial intelligence can pass the Turing test, which means they can be indistinguishable from a real human in a blind test, according to HPC Wire.

See Now: NASA's Juno Spacecraft's Rendezvous With Jupiter's Mammoth Cyclone

©2017 ScienceWorldReport.com All rights reserved. Do not reproduce without permission. The window to the world of science news.

Join the Conversation

Real Time Analytics