Slashdot Mirror
Google Hires Joke Writers From Pixar and The Onion To Make Assistant More Personable (cnet.com)
One of the biggest announcements made at Google I/O earlier this year and at Google's hardware launch event this past week was Google Home, an always-listening wireless speaker that features the Google Assistant. The Google Assistant is similar to Amazon Echo's voice assistant named Alexa, as it can deliver search results, sports scores, calendar information, and a whole lot more. But in an effort to make the Assistant more personable to better compete with Siri, Alexa, and Cortana, Google has decided to hire joke writers from Pixar and The Onion. An anonymous reader quotes CNET: According to a Wall Street Journal report, comedy and joke writers from Pixar movies and the Onion are already working on making Google's upcoming Assistant AI voice service feel more loose and vibrant. The development of compelling voice AI will need to start drawing from deeper, more entertaining wells, especially as these home hubs try to have conversations all day long. Current voice AI like Apple's Siri and Amazon's Alexa on the Echo try to engage with personality, and they even tell jokes (usually, bad ones). But, as these services aim to be entirely voice-based, like the upcoming Google Home hub, they'll need to feel more alive and less canned. Google Home debuts this November, and the upcoming Google Pixel phone, arriving in stores and online on October 20, is the first Google product featuring the new Assistant voice service.
maybe they should personify a paperclip.
Setting aside how funny this is or well it works, it seems a reasonable place to throw a few bucks. While it won't be enough to tip my decisions, I've seen the Cortana easter eggs; despite not being a functionality, flavor can definitely increase a feature's appeal and, perhaps more importantly, branding. The industry is desperate to crack the code that causes nonsense like Over9000 to go pandemic, but this can have a similar effect (at a far tinier, subtler level).
Sure, it can backfire if you half-ass it, but apparently they're serious enough to spend a little money. Whether the hires are well-chosen is left as an exercise for the flamewars.
You make a good point... why are they hiring writers from the Onion when there are thousands of good writers in Hollywood or the video game industry?
The Onion specializes in brutal, bitingly dry satire. Not so great at witty banter and DIALOG. Pretty sure if I ask an "AI" assistant a question I'd prefer a useful, relevant, and personable answer, not a monologue about the banality of middle America.
I want my personal assistant's bon mots to be whoever wrote the filler dialog to The Wire, Archer, Baldur's Gate II, or Mass Effect (among hundreds of other valid examples)...
The very first thing I tried to do with my friend's Echo was to activate the self destruct sequence. It was immensely disappointing to get a "don't understand" response instead of "Does the first officer concur?"
They may be completely useless, but easter eggs make excellent first impressions. You shouldn't necessarily need joke writers for it, though. Easiest way to figure out what to say in response to the user is to simply field test the crap out of it and google any curious-sounding repeated questions that you notice. That, plus maybe you get a meme database and try to include most of the more popular ones. You don't have to spam the user with references; just respond intelligently if they appear to be alluding to one.
with my appliances. That's for some of the people I know, not any machines.
I don't want it cracking jokes some marketer thinks I should hear - I have TV for that.
I don't want it cooing soothing burbles if the "AI" figures I'm not feeling up to snuff. I have friends for that.
I don't want it trying to carry on a faux conversation with me. I get enough of that crap in the real world.
In short, if I at all want a voice interactive machine, I want it to listen to what I command, interpret said command correctly, execute it correctly and then, if warranted, respond with the results.
The headline does sound like an Onion article. But its getting harder and harder to distinguish reality from The Onion.
Come November 9, I may have to seriously consider the possibility that I am living inside a Matrix-y simulation inside an Onion article of an advanced civilisation.
already working on making Google's upcoming Assistant AI voice service feel more loose and vibrant
That sounds like they're making a completely different kind of "assistant."
systemd is Roko's Basilisk.
I don't want a joke unless I specifically as for it. Case in point: last year I was driving from AZ to CA on I-8 and my speedometer cable broke so I couldn't tell how fast I was going. I thought, "My phone has GPS which can give me my speed so I'll ask Siri." "Hey, Siri, how fast am I going?" She responded, "I've been wondering that for a while." Great, thanks for nothing, smartass.