One of the more impressive announcements – and demonstrations – at the Google I/O keynote this week was their Google Duplex calling function that will see the Google Assistant calling businesses and interacting with real people. While impressive, the demo has raised some concerns.
With the Google Duplex demo on-stage at I/O, it was quite apparent that the calls shown were of the highest quality. The person interacting with the Google Assistant seemed unaware that they were speaking with a machine, with the Assistant dropping in human speech affectations like ‘umm’, ‘ah’ and even ‘Mmm-hmm’. This level of speech and the ability to correct throughout the calls showed a level of excellence which would see the Assistant possibly passing the Turing test.
While having an Assistant able to call people and interact with them from a pure technical standpoint is impressive, it’s worrisome for a lot of people. Google is getting ahead of the curve issuing a statement to CNet, saying
We understand and value the discussion around Google Duplex — as we’ve said from the beginning, transparency in the technology is important. We are designing this feature with disclosure built-in, and we’ll make sure the system is appropriately identified. What we showed at I/O was an early technology demo, and we look forward to incorporating feedback as we develop this into a product.
Exactly how Google will move forward with this is part of the Google Duplex ‘experimental’ status. The feature is only launching in the US to start with while they iron out these issues and work out how best to inform a person speaking with the Assistant that they aren’t talking to a real person. We’ll know more once Google begins rolling out Duplex to US users in the ‘coming weeks’.
How do you think Google should approach this?