Tag Archives: Artificial intelligence

Psychology, Tech

Hey, What’s Your Name?

Google Home Mini

Image by antonbe from Pixabay

Our current group (party, gaggle, company) of smart-speaker voice assistants have real names, Siri, Alexa, Cortana and Google.  Eh.  Why didn’t Google give their assistant a name, a proper name.  She clearly has a personality of sorts, even if it is pre-programmed, and apparently now has feelings to hurt if you swear at her or speak impolitely.  If she was a real person she would, I’m sure, be embarrassed by her name and people would say “did her parents have a bad sense of humour?”

It’s not just me being flippant, for me it just doesn’t feel natural saying “OK Google” or “Hey Google”.  I’m sure they think it sounds cool but imagine if everyone who meets you and asks you a question had to prefix it with “Hey Joe” or “Hey Sue” or worse still “Hey Human” it would get boring very quickly, for both parties.  As for “OK Joe” – that just sounds unnecessarily aggressive.  I think it’s the “OK” or “Hey” prefix that niggles, just asking “Google?” would be a little better, to be fair.

If we are to, in the future, have a natural verbal interaction with technology it has to be exactly that – natural – not an excuse to crowbar the name of the service provider into the conversation.  At the end of the day you’re likely to know you’re using a Google device.  Their approach seems to hark back too much to the original Star Trek’s “Computer…”

Hey Google, please grace your AI with a real name, those of us who feel daft saying “Hey Google” might use it more.  Personally, having lived in Newark, and more specifically worked with a certain Mr Johnson for twenty years, a natural way to summon her would be to say “Now then Mush, what’s the weather like today?”

Advertisement
Standard
Tech

Becoming Too Human?

English: The following is the author's descrip...

English: The following is the author’s description of the photograph quoted directly from the photograph’s Flickr page. “Researchers from many fields will use the new IBM Blue Gene/P supercomputer at the Argonne Leadership Computing Facility. Photo, courtesy of Argonne National Laboratory.” (Photo credit: Wikipedia)

It seems that Artificial Intelligence still has some way to go.

Last year a computer was given YouTube to look at and it learned how to identify a cat, now it has been revealed that IBM’s Watson computer which famously won a game of Jeopardy on American TV was given access to Urban Dictionary as part of its education in the English language – to enable it to understand the nuances of the language and slang terms.  Unfortunately, due to the level of fruity language in Urban Dictionary and its inability to distinguish normal from profane language it simply learned how to swear, at one point using the word “bullshit” in answer to a researcher’s question.

As commenter Bleary said on Gizmodo UK it would have changed the film 2001: A Space Odyssey somewhat – “My mind…it’s going….I can fucking feel it.”   “Open the pod bay doors HAL,”  “Fuck you Dave”

Hmm, if the computers can’t decide what’s appropriate in polite society what else could they be capable of?…

[Gizmodo UK]

Just as an aside – as I was choosing the image at the top, of IBM’s Blue Gene supercomputer I realised that David Bowie’s Blue Jean was on the radio.  Spooky.

Standard
Tech

I’m Listening

Quietz! I hearz sumpin coming...

Quietz! I hearz sumpin coming… (Photo credit: pjern)

Voice control of computers has been a dream since before Scotty tried to chat up an Apple Mac in that Star Trek film and now processing power is enabling it to be a reality even though it is still comparatively basic at the moment; even Apple’s Siri is a human-friendly front end of what is effectively a search engine.  Both Siri and Android’s voice actions allow commands to be given to the devices and although they are pretty good at recognising what you ask them to do it’s still not an artificial intelligence.

Nuance, the company that created the technology behind Siri, are working on voice recognition systems that don’t need to be told when to listen (by a tap or a voice command like “Hi Siri”, “Xbox listen” or “Computer?”).  These systems are always listening, just waiting for you to say something that it might be able to do something about; just mumble “I wonder what the weather’s going to be like at the weekend” and your phone will instantly have the weather news for you like the world’s fastest personal assistant, never having to be asked, always ready with the answer.  The idea has great potential in streamlining device use, or customizing the information shown on services like Google Now.

But how annoying could it become if you’re having a normal conversation or even talking to yourself and your phone lights up “sorry, I didn’t catch that, do you want me to find something for you?”  to which you instinctively say “no, I wasn’t talking to you.”  Even more annoying is when your phone replies “oh, well if you’re going to be like that.” and sulks for two days.

No doubt the software will eventually have ways of detecting whether there is more than one voice being heard so it can ignore questions that aren’t directed at it and just sit there making notes about what you and your friend, relative, partner or cat are talking about in case it can find something relevant should it be called upon but there could still be occasions when it may go off and search for something that it shouldn’t perhaps.  Will it apologise for getting you into an embarrassing situation based on something it heard on a tv show?

Of course this will have the conspiracy theorists worried that it’s sending everything you say to the government but that’s inevitable, they probably also think the government’s reading their emails too.  Now where’s my phone hiding?

[Gizmodo UK]

Standard