Google Duplex broke the Turing test and leaps across the uncanny valley

This is the editor’s letter for the May 2018 edition of CanadianCIO.

If we push artificial intelligence across the uncanny valley, what fate awaits us on the other side?

Given what we’re seeing from Google Duplex, we might be very close to finding out. Demonstrated at Google’s IO conference just last week, Duplex is a new natural language processing system that can imitate a person and hold simple conversations for the purpose of booking appointments. It’s a new feature for Google Assistant that the search giant says it will be testing live in the field starting this summer. But it’s also a much more significant moment for computing.

In another talk at Google IO, Alphabet chairman John Hennessy (Alphabet is Google’s parent corporation) said that Duplex passes the Turing test when it comes to the task of making appointments. Developed in 1950 by British computer scientist Alan Turing, the test was designed to see if a human could be fooled into thinking a computer was human.

Blade Runner 2049 - Luv
Hi, I’d like to book a haircut appointment, please. (Photo by Stephen Vaughan – © 2017 Alcon Entertainment, LLC.)

In the samples of Duplex played on stage at Google IO, the algorithm-created voices sounded completely human and natural. There are ‘ums’ and ‘ahs’, and pauses and emphasis in the voice. The humans on the other end of the interaction thought nothing of having the conversation required to schedule an appointment for a haircut.

Even if it’s limited in terms of the scope in passing the Turing test, the fact Duplex tricked human beings is evocative of some sci-fi dystopias we’ve seen depicted in fiction. In Blade Runner, replicants are fabricated organisms that can’t be distinguished from human beings unless they’re subjected to a Voight-Kampff test. In Battlestar Galactica, Cylons are synthetic lifeforms that easily pass as human. In both of these imaginary futures, humans are living a bleak existence or not living at all.

But if the public policy leaders that want to apply AI in a way that’s guided by a set of ethical standards, we might just avoid a future where our own artificial creations seek to annihilate us. The new Government of Canada CIO Alex Benay has reached beyond his own role by helping to form the CIO Strategy Council, where AI ethics were a focus of discussion. The former Information and Privacy Commissioner of Ontario, Ann Cavoukian, is expanding her internationally-lauded Privacy by Design principles to include AI ethics principles.

Already, these public leaders are helping to make those applying AI more conscientious about the risks. When I was speaking with City of Mississauga CIO Shawn Slack this week, he mentioned that he’d seen Cavoukian’s ethics principles and planned to use them as a guiding light as his team adopted AI.

If all CIOs are equally as moral as their organizations adopt AI, then we might just have a chance against the robots. Then when we find ourselves on the other side of the uncanny valley, we’ll enjoy green pastures rather than a silicon wasteland.

Would you recommend this article?


Thanks for taking the time to let us know what you think of this article!
We'd love to hear your opinion about this or any other story you read in our publication.

Jim Love, Chief Content Officer, IT World Canada

Featured Download

Brian Jackson
Brian Jackson
Former editorial director of IT World Canada. Current research director at Info-Tech

Featured Articles

Cybersecurity in 2024: Priorities and challenges for Canadian organizations 

By Derek Manky As predictions for 2024 point to the continued expansion...

Survey shows generative AI is a top priority for Canadian corporate leaders.

Leaders are devoting significant budget to generative AI for 2024 Canadian corporate...

Related Tech News

Tech Jobs

Our experienced team of journalists and bloggers bring you engaging in-depth interviews, videos and content targeted to IT professionals and line-of-business executives.

Tech Companies Hiring Right Now