It looks like magic.
Let’s face it, the majority of people do not understand how this technology works. Nor do they really need to, but the audio calls sound so realistic it has resulted in a mix of both amazement and fear as well as claims that it is morally wrong to mimic a human.
I am surprised Google did not address this in the presentation but it did react quickly after it started to gain currency online. Falling back on the technology’s early state as an excuse and saying it will evolve, Google emphasised that the system would identify itself in some way so that people were aware they were interacting with technology rather than a human.
Was that a last ditch effort to stem the tide or was that the intention all along? I suspect the former. The way people will respond is going to completely change once they know it is a robot but I think they realised there is no other way to begin with. Eventually though, it will become the norm and self identification will be unnecessary.
The mitigating factor, that may save it from failure, is that this is being used in a business setting. It is a brave business that forgoes a client because they decided to have some fun trying to trip over the technology.
Regardless, many will try to do so in the early days and the question is whether the technology can get through that phase and survive. Either way, the underlying capability shows how realistic robots will become.
For me, I don’t want to have staccatoed conversations with technology - and the only way to have a conversation is in the language I understand and feel comfortable with so to suggest this is a bad move by Google is I think wrong.