Thursday, January 23, 2020

Google Duplex: Using A.I. to Mimic Humans

Every year, Google holds its own developers conference to announce its latest and greatest software and hardware that will be soon released to the public. Google I/O, as they call it, stands for "Innovation in the Open". Since 2008, these conferences have been more or less the same, including updates to their current software such as Android, Google Drive, Chrome, Gmail, and new products such as Google Home and Google Pixel.

However, in 2018, Google announced something truly remarkable. They called it "Google Duplex", an A.I. assistant for users to use that can make appointments or purchases with local businesses. Take a look at their quick demonstration below.


If you have ever used a mobile assistant such as Siri, Alexa, or even Google Assistant, it's fairly easy to distinguish them from a real human due to their inability to carry on a conversation, speech patterns, and general sound of their voice. Google Duplex, however, takes it to another level. In both examples, it was talking to a real human, yet they couldn't tell at all that they were talking to a robot. The sound quality of the second phone call wasn't the best and the lady even had an accent, yet the A.I. was able to continue the conversation perfectly. The fact that Google programmed it to use slight pauses and filler words in between sentences in order to mimic a real person is simply amazing and terrifying at the same time. 

Based on James H. Moor's three stages of a technological revolution, Google Duplex is currently in the permeation stage as only certain Pixel devices are able to use it and the only tasks it is able to complete are restaurant reservations and purchasing movie tickets. It is easy to see how incredibly useful Google Duplex can be, however, Moor has stated that "inevitably revolutionary technology will provide numerous novel opportunities for action for which well thought out ethical policies will not have been developed". The current ethical issue in hand is that does Google have any obligation to tell someone they are talking to a robot? How would people react on the other end? It's inevitable that better ethics will be needed as this level of technology becomes more than just a friendly helper on your mobile phone. 

1 comment:

  1. Hi Steven, I think you chose a very interesting topic to blog about. In particular, the ethical questions regarding the development of AI and the relationship between big tech companies like Google and their customers are questions that must be answered. Like James H. Moor said - We should be more proactive in regards to ethics rather than reactive. However, I believe you can improve this blog by shaping your blog around the readings as opposed to mentioning just the connection at the end.

    ReplyDelete

Note: Only a member of this blog may post a comment.