In 100 Years AI Will Be _____
Candid, Relational, Natural, and… Transparent?
In this short, post, I discuss some of my hopes, fears, and predictions of voice assistant AI.

On May 9, 2018, Google released a new virtual assistant feature called Duplex. Unlike the typical bot voice we have come to expect from Google Assistant and most other products of its kind, Duplex surprised Google I/O conference attendees with a convincingly human voice. This voice was indistinguishable as a robot built upon natural language processing and AI components. To demonstrate its ability to “pass” as a human, Duplex performed a live phone call to a hair stylist. The voice — a slightly airy and cracked voice of a young professional woman — managed to book a haircut over the phone by navigating a complex conversation, complete with umms, ahhs, and unexpected changes in scheduling.
Many in the crowd were delighted by the bot’s candor. “At last” a virtual assistant that sounds “good.” To me, this moment was a glimpse into what the future of virtual assistant application of AI will look like: candid, relational, and no longer synthetic. But it was missing one major piece I had hoped for: transparent. The robotic voice typical of Amazon’s Alexa, Apple’s Siri, Google Assistant, and Microsoft’s Cortana are currently stilted, awkward, and undeniably machine fabricated in their current state. This has been a pain point that has turned away potential technology adopters, but it has also reassured those who want to maintain clarity that they are talking to robots. In many ways, the synthetic voice has kept users from falling into the Uncanny Valley of the Mind.
We can not know whether Google’s intent was to step over current AI ethical standards or to conduct a social experiment on an unknowing audience — but we can discern from the display of Duplex that humans appear delighted by robots with human voices. Natural language processing and digitally generated conversation exists as a technology that will soon be ready for a multitude of products and we need to understand how humans process and react to such a feature.
My main concern for the future of AI is that we will continue to create new products “just because we can,” without first establishing the problems we are are solving and the problems we are potentially creating. In short, we are kids playing in a sandbox and our actions have the potential to impact everyone in our global society.
My hope for the near future is that product designers will quickly find an ethical medium between highly relational AI voice assistants and the control humans want to maintain over their lives, their data, their privacy, and their relationships. After the tidal wave of concerns voiced to Google about Duplex, a disclaimer was added to the product. Additionally, Duplex will now only be rolled out to participating businesses. These are good standards, but they should have come much earlier in the product design.

