Google Duplex – Another Breakthrough in AI World. Is It Good Or Bad For Us?
Earlier in May 2018, Google CEO Sundar Pichai disclosed another jaw-dropping addition at Google’s annual I/O developer conference— the Google Duplex. Used with Google Assistance, the Duplex can make essential phone calls on your behalf, mimicking your voice!
After the successful demonstration of the new feature, Google Duplex now becomes a hot topic for discussion.
Overcoming the issues of the previous voice assistant apps, Google duplex, the AI-based technology, brings a great revolution in generating the natural conversations to replace the phone call tasks.
What is Google Duplex? How is it working to aid the users? What are the features and lacking? Let’s check this out.
Google Duplex – A Robust AI Phone Assistant!
Many technologies have been launched targeting the goal of having natural conversations with humans with the computers as well. However, all these technologies are unable to recognize the simple words and commands. It has become irritating to interact with the artificial computerized voices.
With the announcement of the jaw-dropping and mind-blowing technology, the Google Duplex, many of the issues are now supposed to be solved. Employing AI, Google Duplex will manage all conversations in natural sounds. Unlike Google Home Speaker, that sounds semi-robotic, the incredible Google Duplex’s system bears “speech disfluencies” like ‘hmm’ and ‘uh’. Such calibration let the Google Duplex sound like a real person with appropriate tempo, pauses, and word elongation as if someone is thinking.
The remarkable technology can replace several ‘real phone’ tasks like scheduling appointments, reserving tables in restaurants, etc. The system makes the entire conversation so natural that no one can guess that on the other side of the phone there is a machine, not a human!
What Google Duplex Can Do For Us
Google Duplex will supposedly conduct the conversations naturally and comfortably. Once correctly and intensely trained in closed domains, Google Duplex is ready to give you a natural talk. Due to the astonishing attributes, it is enough to replace the human need for various real-world phone engagements. From the business affairs to the daily routine work, Google Duplex is posing an enormous impact in different life fields.
The developers actually utilized real-time supervised training to guide Google Duplex in a wide range of domains. Such profound training makes the system unbeatable by any of the voice assistance software.
Let’s have a look how the Google Duplex is benefiting the humans.
RNN — To Conduct Natural Conversations
Usually, when we talk with one another, we sometimes use complex sentences. Many times we correct ourselves in the mid of the sentences. Depending on the context, the same sentence or phrase can have several meanings. Additionally, during a spontaneous conversation, people speak faster and less precise. This all causes the natural language difficult to understand. Therefore, the speech recognition by any ordinary machine is harder.
Fortunately, with advance Understanding, Interacting, Timing, and Responding, Google Duplex offers the most natural conversation you can ever have.
It provides high precision due to extensive training on a vast quantity of anonymous phone conversation data. Google Duplex has a recurrent neural network (RNN), which is constructed using TensorFlow Extended (TFX). RNN employs Google’s automatic speech recognition (ASR) technology outputs and information from the audio. Finally, Google Duplex makes use of the hyper parameter optimization from TFX. Such deep training and features are enough to cope up with the natural language challenges.
Sounding Natural via TTS Engines and Speech Disfluencies
To produce a natural sound, similar to the human voice, Google Duplex is accompanied by a combination of two text-to-speech engines- a concatenative and a synthesis TTS engine. These are responsible for controlling pitch in accordance with the situation.
For more natural and familiar touch, Google Duplex employs Speech Disfluencies (e.g., hmm and uh). The disfluencies are added at the time of combining different sound units in concatenative TTS engine. Sometimes synthetic waits are also inserted. Both of these are responsible for letting the system signal in a natural manner while processing. This resembles as if someone is gathering thoughts before answering.
Matching Latency Expectations
Meeting people’s latency expectation is also a significant challenge. Google Duplex is well aware of responding back within the expected time. When required low latency, it uses faster and low-confidence speech recognition models. To show a person hesitance while responding, Google Duplex utilizes faster approximations without waiting for RNN to work first. Likewise, while replying to a complex sentence or phrase, more latency is introduced. These all features make the conversation less robotic and more natural.
Self-monitoring
Usually, Google Duplex system can automatically carry out refined conversations. Since it is based on Artificial Intelligence, therefore no human involvement is required.
The remarkable feature of the Google Duplex is its ability of self-monitoring. It can detect whether it can handle the task by itself or not. If the Google Duplex feels that it cannot tackle the situation autonomously, then the system signals the task to the Human operator without any delay.
Helpful For Business Organizations and Users
Google Duplex launch dramatically influences businesses tasks as well as aids the users in many aspects.
Considering the business engagements first, we can see that Google Duplex is playing a pronounced role in supporting appointment based business activities. It allows the customer to book via Google Assistance, with no risk of date and time alteration. It also reminds the customer about their upcoming appointments. Hence, one can easily manage any amendments, like cancellation or rescheduling, through this technology.
Google Duplex helps the organizations by reducing unimportant phone calls by the customers like for inquiring working hours or holiday.
Not only the in the corporate world, but Google Duplex is also applicable and suitable for the daily affairs as well. The user can easily avoid phone call and simply interact with the Google Assistance. Now, without involving the user, Google Duplex will take over the rest and make a call on his/her behalf.
Google Duplex is the great invention for the hearing-impaired users as well. Eliminating the language barriers and address accessibility, the impaired users can carry out their telephonic tasks by using Google Duplex.
No Need Of Internet
Many of the software used for the assistance of the users are mostly works with the internet only. You will be shocked to know that you don’t need to connect to the internet first for using Google Duplex.
Google Duplex is the offline service of the Google Assistance. This makes the system easily accessible and free for any user. You can schedule appointments or book for the reservations from any place with or without internet.
Is It Really That Worthy? Downsides Of Google Duplex
Google Duplex has been unveiled with a successful demo of previously recorded phone calls between a human and a computer. The examples have left the audience jaw-dropped, as you cannot identify the computer’s voice. With the aid of AI, Google Duplex offers contextual and intensely trained conversation in the most natural style.
The presentation about the system was so impressive and stunning that no one could deny. It was the first flawless experience of human sounding AI interaction with the real human.
The outstanding voice of the computer has led to so many queries and apprehensions as well. The human-like computerized sound with human-like pitch and pauses can easily be used negatively as well. Therefore, despite the overwhelming appreciation, a debate has now started about the system.
The accurate mimicry of the human sound by the AI-based system and easy access by anybody has left many frightened.
- The fabulous properties of the Google Duplex can be used to make people fool as well.
- Since this technology requires no internet connection, it can be a handy tool for the terrorists. They can easily interact with one another and can also make fake calls to divert the attention of police.
- People can use it for fun purpose too like making phone calls to salons and restaurants just for asking about the false appointment.
- The technology also possesses interesting features for telemarketing purposes, where it may advertise the products that are injurious to health.
- Kidnappers can also blackmail the person’s family, without having any fear of phone trapping.
- Since it is not the actual voice of the person using Google Duplex, therefore, a person can hide his/her original identity. He/she can use it for any purpose.
Google Duplex Needs To Disclose Its Identity
Apart from the concerns mentioned above regarding the abuse of the technology, there is an ethical issue as well. Google Duplex is somehow breaking the trust of the persons who think they are talking to a person. Many critics have suggested that Google duplex should disclose its identity.
Arvin Narayanan from Princeton University suggested:
“We need ground rules. One proposal: “Turing Red Flag law” — bots should be designed so that it is clear they are bots; they should also identify themselves as bots up front”.
Similarly, Erik Brynjolfsson, an MIT professor, suggested in an interview with the Washington Post: “Instead, AI researchers should make it as easy as possible for humans to tell whether they are interacting with another human or with a machine.”
Therefore, we can say that designing this technology in a way that it first declares itself may possible reduce the wrong usage.
In response to all that Google’ authorities have now declared that they will shortly introduce disclosure feature in the software. Until then we cannot trust Google Duplex blindly.
Do We Have An Alternative?
Due to the risks of fake phone calls via Google Duplex and negative usage of the technology, many are people are hesitant about using it. Though such a fantastic technology is advantageous in many aspects, its ability to completely mimic human voice has left many unnerved. From the ethical point of view as well, this software is lacking yet. Talking to a computer instead of the human representative without any acknowledgment can also distract customers. It can decrease the reputation of the organization too.
Therefore, people from the corporate sector mainly, are seeking for the alternatives, which not only help in business affairs but also attract the customers.
Here we are giving you a brief overview about the best possible alternative – the RosponseAI.
RosponseAI, just like Google Duplex, is an innovative addition to the world of telemarketing. It is also based on Artificial Intelligence, working automatically without human involvement.
RosponseAI, the smart software, is successfully used to replace call center department from the organizations. It is different from Google Duplex, as it is not mimicking human voice. Hence, people will know that they are talking to a computer. RosponseAI only helps the growing organizations for gaining popularity within less time.
Mostly, the organizations are using it because:
- There will be no need to establish a separate Customer Care department and waste budget on hiring employees and on other equipment.
- It is a cost-effective way to produce more live leads and generate profit.
- RosponseAI can replace Customer care representatives’ responsibilities by providing quick automated and relevant responses to the queries.
Hence, we can say that RosponseAI is the best option for the organizations who want to obey the ethical values while still using the latest technology.
Conclusion
Google Duplex is the most incredible invention that will suffice to meet the goal of natural computer-human interaction. It is the fascinating AI-based software that has gained popularity within just a few days of its disclosure. With the use of Google Duplex, computer calls will not sound robotic anymore. No one will catch whether it is a human or a machine to whom you are talking to.
Google Duplex launch has actually opened the window to the near future where technology will be more automatize requiring less human interference. Exciting, beneficial, and a little confronting and creepy, Google Duplex is the best development of the time. Its ability to produce and mimic human voice with appropriate use of verbal pauses in undeniable. Yet there are some drawbacks too. Nevertheless, if the developers seriously look into the matter and solve the issues, then this technology will become unbeatable.