We believe it or not, but it happened on last Tuesday 8th May 2018, when Google CEO Sundar Pichai demonstrated the new technology, called Duplex, during the Company's annual conference for software developers. The Duplex is an artificially intelligence assistant that can call people up and interact with them - without anyone ever knowing they were speaking to a robot.
In his demonstration speech, Pichai has accounted that 60% of US small and medium enterprises do not have digital booking system. He emphasizes this Google Assistant will help them immensely.
When we talk about Artificial Intelligence (AI), we think and compare AI with the expression of human knowledge and emotions in a given circumstance. But this Google Assistant while talking to a human, deployed pauses and sound "ums", "hmm", "I gotcha" more like a human.
Google confirms, Duplex is able to understand complex sentences, fast speech and long remarks and so naturally converse and make an appointment, before sending the user a notification to confirm the booking.
Pichai added
We really want to work hard to get this right.
Paradigm Shift
The minds behind Google Duplex are in the process of changing that paradigm, for better or worse. In the first demo, a woman calls a hair saloon, where another woman answers the phone ; the two go back and forth for approximately one minute, before they figure out a time that works for a hair cutting appointment . In the second demo the assistant pivots to make a new request when it does not get the reasonable information.
The Technology
Duplex was first launched as an experiment few years before and was started by principal engineer Yaniv Leviathan and Yossi Matiass, vice president of engineering . Duplex brings together natural language processing, deep learning and text-to speech technology into one service. The part that resonates most, though, is the "natural" bit—the engineers have trained the Duplex model to match expectations around latency, like pauses after someone says “Hello?”, and to change intonation depending on how the conversation flows. In other words, to react the way humans do when speaking on the phone.
New Challenge
Now the challenge is whether and how to alert people who may not know they are talking to a robot ?
The world is clearly divided. Matthew Fenech, who researches the policy implications of artificial intelligence (AI) for the London based organization Future Advocacy says
That's very impressive, but it can clearly lead to more sinister uses of this type of technology. The ability to pick up on nuance, the human uses additional small phrases - these sorts of cues are very human, and clearly the person on the other end didn't know.
Fenech also pointed out that it is possible to identify nefarious uses of similar kind chat bots like spamming businesses, scamming seniors or making malicious calls using someone's voice.
"You can have potentially very destabilizing situations where people are reported as saying something they never said
Read more at: https://phys.org/news/2018-05-robots-humans.html#jCp
Read more at: https://phys.org/news/2018-05-robots-humans.html#jCp
He also said
You can have potentially very destabilizing situations where people are reported as saying something they never said
Although Google tried to emphasize that the technology is still in experimental stage and not available for consumer devices and it will be rolled out very cautiously when it will fit to use.
Google engineers Yaniv and Yossi wrote in blog-post yesterday,
It's important to us that users and businesses have a good experience with this service, and transparency is a key part of that
Read more at: https://phys.org/news/2018-05-robots-humans.html#jCp
Read more at: https://phys.org/news/2018-05-robots-humans.html#jCp
It is important to us that users and businesses have a good experience with this service and transparency is the key part of that. We want to be clear about the intent of the call so businesses understand the context. we will be experimenting with the right approach over the coming months.
This is the 4 minutes video which demonstrates Duplex
It is still not known how Duplex Technology will address the issues like telecommunications laws which varies globally.
Various ethical and practical questions like consent for call recording, blocking of unsolicited calls etc. etc. are obviously likely to come over the implementation of Duplex for B2B or B2C relationships.
Chris Messina - the product designer , famous for inventing hashtag during his time at Twitter - described the Duplex Technology as
the most incredible, terrifying thing
Google also demonstrated how it was developing tools that could construct whole bots out of people's voices. It showed John Legend recording just a few choice phrases, for instance - which were then synthesized using artificial intelligence into a voice that could be made to say anything and still, sound like the singer.
It is clearly understandable that when Google voice calls us using Duplex could have any voice - making it hard to recognize when it is not a real person.
Conclusion
As of now, Duplex is intended to handle rote responses, like asking to talk to a representative or simple formulaic social interaction. Even so , the programmed Duplex's capability to handle confusion (as on the second demo) is still a significant step forward. With the improvement of AI , voice quality will improve and it will become better at answering more types of questions.
We are obviously still a long way from creating a conscious AI, but we are getting better at the tasks our systems can handle -- and faster than many would've thought possible.
Update
June 26,2018
The Search Giant confirms its plan for a rollout of Duplex after carefully addressing the concerns and worries that they had received in the form of feed back against the initial demo in May 2018. Please read the full report Google prepping its Duplex bot for a summer rollout.
Thank you Tania
ReplyDelete