From suggestive search to voice-enabled personal assistants, futuristic robotics and more, AI is gaining in popularity, with roll-outs seemingly happening every other day. These advances are incredibly exciting, yet it still fills many of us with a sense of unease. For some, it feels as if we’re entering an all too familiar Black Mirror storyline, where AI is furthered beyond our control. And it seems like Google are leading the charge with their new tech ‘Duplex’.
Announced at the Google I/O Conference at the start of the month, Google Duplex is the latest development from the tech giant on the AI front. Developed specifically to sound like a real human voice: pauses, ticks, hesitation and all, this AI tool can simulate real human conversion so well that it easily fooled those who got to hear it before it was announced at the conference. Take a look at the clip below, you won’t believe your ears.
Yes, the man making that reservation is the Google Duplex (not the lady getting everything incorrect). Tech that talks is no new thing, with Siri, Cortana, Google Assistant and Amazon Alexa, we’re all totally familiar with this concept by now. There are plans to potentially bring Google Duplex and integrate it with the rest of the Google products, upgrading the current voice from robotic to… something else. The Google Duplex voice is littered with everyday things that make human speech so distinguishable. Uhm’s and Ah’s, hesitation and inflexion and common features in almost everyone’s conversation, they create the illusion of thought and processing of information, something that pushes beyond most current robotic AI voices. The Duplex voice is bizarrely human, presumably to make it the ‘perfect assistant’. Nick Fox, vice president of product and design for Google Assistant and Search said:
"We think of it as, 'What's a perfect assistant?' When I want an assistant with me, it's there. It's always ready to help. I don't worry about 'Can my assistant do it or not?’ My assistant can do anything I want it to do."
Amazon was the first system like this of note, with their Alexa software integrated with the Amazon Echo in 2014. Since Google entered the game, the public battle is own, with both companies frantically trying to outdo each other in a bid to be the next to market with that special something. Despite the fact that they both have elements that the other doesn’t (you can read about Assistant vs Alexa here), with Google Duplex, it seems like they’ve just done one big leap ahead.
However, the impressiveness of this recorded demo doesn’t a groundbreaking invention make. In actuality, Google is notorious for impressive demos with little to no follow through. In 2017, we were all astonished by Google’s ‘Pixel Buds’, an invention that promised to translate language in real time through an earpiece allowing bilingual ease (kind of) and just five years before that we saw ‘Google Glass’ for the first time. Everyone’s jaws dropped at the prospect of useable, functional AR, however, the design was unattractive (cycling glasses are bad enough on cyclists) and any notion of its success frittered away to nothing, ending with Google announcing the ceasefire of their Google Glass dreams.
Right now, the Duplex is still in the creation phase, with requests for a live demo being denied, however, it does seem like the company are preparing for a big move. Last year, they introduced a wide price range of home speakers, from the Google Mini ($40) to the Home Max ($400), they are also looking for these products to be available in 80 countries and 30 languages by the end of the year and Google said the Assistant now partners with 5,000 household connected devices, which is more than triple of that from January. The rivalry lives on through, with Amazon being a little further ahead in these elements, however, with the implementation of tech that can not only simulate human speech but also order, book appointments and chat to service staff on your behalf, Amazon could be about to get blown out of the water.
Moving on to more serious and philosophical matters, what does the use of life-like AI mean for us? And what does the increasing use of AI technology mean for security and privacy? From movies like ‘Her’ to TV shows like Channel 4’s ‘Humans’, AI and the use of it is questioned regularly and often feared. Generally, it all seems like a good idea until one of them starts freaking out, laughing like a Disney Villain and taking over the world. Then we all panic and wonder what it is that’s actually stored inside those tiny little speakers, not to mention what they can hear…
Following on from the Facebook Analytica scandal and the imminent GDPR changes, concern about privacy and security is at an all-time high. While the interrogation light is shining brightly in the eye of Mark Zuckerberg at the moment, few people are remaining aware of what Google knows about them. While you tell Facebook what you want it to know about you, Google knows everything else: what you search for, where you go, where you review, often where you are and much much more. Their knowledge about you is pretty much endless, and it was skirted around at the I/O conference well, just take a look at this article from Huffington Post about a Twitter thread that revealed it all. If Google can trace everything about anyone who uses their search engine and products, who’s to say that they don’t have the technology to listen through their Assistants and into your home? We know we sound like we’ve glued our tinfoil hats to our heads, but combined with this incredibly human AI (should it be successful and rolled out), it’s another ethical question that needs to be considered.
When you listened to the recording above were you aware that it was an AI? Think about if you didn’t know, would you be able to tell? Probably not. In terms of this, should AI have to disclose that they are AI before the conversation proceeds? Not to mention, if this ends up being successful, it means that we’re looking at a lot more products (outside of home systems) that will be able to utilise it. Looking at the way humanity is with technology and expansion, it’s likely that even more work will go into AI and its advancement, perhaps one day nobody will be able to tell the difference between human and system. Going back to Google Duplex, Fox admitted that the Google team are still learning about this and will continue to learn as Duplex is developed. He said:
"We need to set the right expectations with everyone, these are implementation questions of the technology that I'd say, humbly, we don't quite know all the answers to yet, and need to be figured out as we see this in the real world."
This technology has the opportunity to swing both ways, there is the human element of the AI voice, which could make the tech seem more approachable and has some potential immediate benefits those leading isolated lives, like a large proportion of the elderly population. It really is eerily lifelike and could act as a voice to talk to for those who might not have it. In fact, Google is one step ahead of this, they are already giving their Assistants pre-recorded human voices, including that of singer John Legend. Making him the first celebrity voice to be widely and commercially available, outside of things like children’s toys and comedy sat navs. However, the fact that it is likely to pass the Turing Test may have people worried about its potential to learn a little more than it’s supposed to.
“The development of full artificial intelligence could spell the end of the human race….It would take off on its own, and re-design itself at an ever-increasing rate. Humans, who are limited by slow biological evolution, couldn't compete, and would be superseded.”
— Stephen Hawking
Let’s just hope that the wonderful Hawking was incorrect in his above prediction to the BBC.
We won’t know what will happen until it happens, and there’s a little while yet until Google Duplex can be snapped up by the masses but we will leave you with a question: if it were available today, would you buy a Google Duplex?