A long-standing objective of human-computer interaction has been to empower users to have a natural conversation with computers, as they would with each other.
In these years, we have seen the ability of computers to understand and to improve normal speech, in particular with the application of deep neural networks (e.g., Google voice search, WaveNet).
Still, even with the present state of the art systems, it is frequently frustrating talking to computerized voices that don’t understand natural language.
In particular, automated phone systems are as yet struggling to recognize simple words and orders. They don’t take part in a conversation flow and force the caller to adjust to the system instead of the system adjusting to the caller.
At last, artificial intelligence is being put to good use: making phone calls.
Google revealed Duplex last week, an innovation for conducting organic conversations to carry out particular tasks over the phone.
You cannot use Duplex to call your mother once a week. But if you need to schedule an appointment or book a reservation, Duplex is your man … er, software.
Not at all like the other emotionless voice assistants we’ve come to know and love, Duplex sounds natural; the system barely shows its robotic side, speaking as smoothly as anyone with a firm grasp of the English language.
What Is Google Duplex?
Google Duplex technology is created to sound normal, to make the conversation experience more friendly. It’s important to us that clients and businesses have a decent experience with this service, and transparency is a key piece of that. We need to be clear about the purpose of the call so organizations understand the context.
What will I use Google Duplex for?
Google Duplex won’t say whatever you instruct it to — it only works for specific kinds of over-the-phone requests. The two illustrations Google introduced at I/O involved setting up a hair salon appointment and a reservation at a restaurant. Another illustration is getting some information about business hours. As indicated by a broad post on the innovation on Google’s AI Blog, Duplex has been purposefully constrained to “closed domains.” In other words, it only works for those scenarios because it’s been trained for them, and cannot speak freely in a general context.
How does Google Duplex work?
Like a majority of Google’s most inventive innovation, Duplex relies upon a machine learning model drawn from real-world data — for this circumstance, phone conversations. It consolidates the organization’s most recent leaps forward in speech recognition and content to-speech synthesis with a range of contextual details, including the purpose and history of the conversation.
To make Duplex sound normal, Google has even attempted to reproduce the flaw of common human speech. Duplex consolidates a short delay before issuing certain responses and says “uh” and “well” as it artificially reviews data. The outcome is phenomenally exact with amazingly short latency.
When will Google Duplex be ready to use?
Google duplex will be released by mid-year, with testing focusing on restaurant reservations, hair salon appointments and asking about holiday hours. It’s unclear if Duplex will be restricted to those task upon launch, yet it’s a safe bet that those three things will be a focal point of the new feature if that is the thing that Google intends to test.
What are the limitations of Google Duplex?
As normal as Duplex sounds — at least in these initial demos — there seem to be limited to what the AI can do now. Google says Duplex “is equipped for doing modern conversations and it finishes the larger part of its tasks fully autonomously, but the keyword there is a majority.” Based on what we’ve seen, Duplex works best when it’s performing a very specific task and less suited to open-ended tasks like follow-up inquiries on specific services.
Allowing users to interact with technology as normally as they interact with each other has been a long-standing promise. Google Duplex makes a stride toward this path, making interactions with technology by means of regular conversation a reality in specific scenarios. We can accept the fact that these innovations and technology advances will ultimately contribute to a meaningful improvement in user’s experience in day-to-day interactions with computers.