The reactions yesterday when Google unveiled its Duplex AI natural conversation technology ran the gamut from amazement to fear of impending robotic doom. Duplex is what I call iAI – or interactive AI –and it’s clearly moving at a dizzying pace. However, the future of iAI is dependent on incorporating decision making and guidance capabilities into these systems, and when that happens, it will be a game-changer.
First, a re-cap: At its I/O conference, Google CEO Sundar Pichai played phone calls placed by Google Assistant to a hair salon and a restaurant – the type of simple task-oriented conversations we all need to do everyday but don’t always have time for. The voices sounded uncannily human, and the interactions, even when there were misunderstandings, seemed authentic. You can watch video and hear those calls here.
Lots of people were excited about the possibilities of Duplex. Some people expressed concern. Here’s how Mark Bergen at Bloomberg described it:
“Outside the Google technology bubble, critics pounced. The company is placing robots in conversations with humans, without those people realizing. The obvious question soon followed: Should AI software that’s smart enough to trick humans be forced to disclose itself. Google executives don’t have a clear answer yet. Duplex emerged at a sensitive time for technology companies, and the feature hasn’t helped alleviate questions about their growing power over data, automation software and the consequences for privacy and work.”
These are important issues to ponder, and there’s no doubt Google is grappling with these questions internally. The technology is in development and the system still being trained for particular domains and disciplines. It looks like Google intends to employ Duplex to place calls to small businesses for information on things like holiday weekend hours, testing the system while updating its business data at the same time.
We can expect interactions such as those displayed by Pichai yesterday to become commonplace and increasingly competent as iAI technologies improve. But for that to happen, there’s an imperative element that must be present for any of these systems to become truly reliable iAI, and that’s the element of decision-making.
Consider the hair salon scenario: The Duplex system has a simple task of finding a salon appointment for a women’s haircut within a specific timeframe on a specific date. It is able to parse its available data to determine that a 10am appointment would work. Yet, what if it could go further, perhaps inquiring about the other types of services available the same way a real-live assistant might, knowing that her client appreciates a scalp and shoulder massage if it’s available and she has time – ? This is the difference between the simpler task-oriented state of what Google introduced and what true decision-making capabilities can bring.
The first company that can develop iAI technology that can partner with a human to assist in decision making and instill confidence in that decision will be dominant in this race. How is it possible? Ideally, along with a corpus of decision-specific data, the technology would require a framework to define the decision, and would need to know something about the person with whom and on behalf of whom it is interacting.
This decision-making framework and capability is precisely what comprises the core value of what we are building today at Element Data with Decision Cloud. I’m excited about what’s on the horizon.
–Chuck Davis, Chief Technology Officer, Element Data Inc.