This is one category of bug that's always intrigued me - I can understand unintentional triggers of Alexa occurring (as they did with the Home mini at launch), but why are those triggers resulting in a "creepy laugh"? Is it just that we're only hearing about the unintentional triggers that result in a laugh?
My Echo has triggered randomly when I'm nearbye and not talking. Only to say something to the effect of "Sorry I didn't understand what you say" so I'm guessing there is a bit of a bias in people reporting.
I will be honest with you. Lately everyone's been grinding the word "terrifying" into the carpet with their boots. But this. I was simply not prepared for this. This was terrifying.
It feels too specifically creepy to be a random bug?
My guess is that a developer replaced the "I didn't hear what you said." response with this creepy laugh on his home Alexa, then somehow accidentally pushed that build to the public.
The other explanations are a hack or disgruntled employee which seem worse.
I've never heard ours laugh in any context. Rarely it will trigger seemingly without provocation, but in that case it says it couldn't understand a question, or does something silly like define "fire". When it does that, we laugh at it!
The updated article says that the phrase being triggered is "Alexa, laugh". As long as it somehow mishears Alexa, 'laugh' is a very short phrase that seems easy to accidentally detect (maybe even by a cough). That's probably why this one was occurring more frequently than other possible triggers, and would also explain Amazon's fix of changing the trigger phrase to "Alexa, can you laugh?".
The command is "Alexa, laugh". Audio-wise, a mere cough could be interpreted as "laugh", that's why Amazon's solution is to change the command into something more pronounced.
Other unprompted activations have been reported in abundance when Amazon installed chatbot functionality after the Alexa Prize: Alexa would start having conversations with herself in the middle of the night. One can expect bugs for newly added features, and of course the creepy ones are more likely to make the news than any random glitch.
But I wonder, can their response be trusted? I mean, it's either malice or incompetence or a complete fluke accident, and the third one is the only one they'll admit to, right?
Malice is the first thought that comes to my mind. Perhaps a security vulnerability allowed somebody to access people's devices. It sounds exactly like the type of joke a curious hacker would play.
If it's a result of something shady they're doing, why would they release a post mortem instead of just letting the story die and let people forget about it?
Wow, you know, I never really thought about it, but this highlights the political premise that all postmortems are only expressions of either extreme shame or heroism, and perhaps only rarely honestly recount mediocre transgressions.
Yes, I think this is just too dystopian for rational people to be allowing to occur. We've actually made intrusive, spying devices which laugh at their users. (Note, they're not owners, but just users of a service owned by a mega-corp.)
I think anyone who thinks that this is okay, didn't have their prescribed dosage of Max Headroom re-runs.
It is just too dystopian to be real. We can't seriously be running as fast as we can off this cliff, really people?
I don't doubt its real. I question the fact that we've ended up in this situation: spying devices laughing at their users.
And we got here by way of technological apathy towards the ethics issues. And .. We are still here. Everyone thinks its okay, because "entrepreneurial triumph", and so on..
We need to step back from this abyss and take a good hard look at what we're making.
I don't understand why many people keep acting like these devices are some sort of sinister spying device that's far worse than anything imagined.
Yet we all carry a far more capable mobile spying device (with audio recording, video+photo capture, and even GPS tracking all built-in) in our pockets and keep it within arms reach at all times.
At least with an Echo you can look at the traffic volume it sends through your Wifi connection to get some sense of when it's sending audio back home but you can't easily audit the traffic your phone sends, you only know what the phone and your carrier report to you.
At least with the Echo, I can be pretty sure that it's only running Amazon software, while that free game I just downloaded might be taking a photo of me right now.
Your point of view is entirely the reason why I'm so upset at these spying devices.
Because you have reasoned that it is acceptable to not have control over who/what observes your life, in the pursuit of .. convenience and leisure ..
Yes, we have lost control over our platforms. Some of us remember a time when it really was the user in control. However, there are now generations of users who simply don't care, they just want the fancy new shit that means they don't have to get off the couch.
And, so on.
(Plus, please see the aforementioned Max Headrom, by way of levity...)
> Yet we all carry a far more capable mobile spying device (with audio recording, video+photo capture, and even GPS tracking all built-in) in our pockets and keep it within arms reach at all times.
I expect it is something that Alexa "hears" and that is the specific response. If you say "Alexa laugh" is that the exact sound that comes back? I will check when I get home. Mine is always misunderstanding what I say. So a false activation + misunderstanding would explain this.
Also, article says it happens often as a bizarre response to requests to turn of the lights. Maybe "lights" is being interpreted as "laugh" as in a command.
This reminds me of the hilarious Alexa AI horror shirt story someone posted in the comments here when Alexa was launched.. can’t seem to find it but it was absolutely hilarious :)
I can't remember where I read it, but I remember a similar article mentioning a very interesting side effect of having decently high-fidelity microphones in these devices. They would pick up noise outside of the range of human hearing and interpret portions of it as a command. It was akin to those articles about the sort of strange phenomenon in machine learning in which computers interpret input wildly differently than humans would due to the underlying modeling the systems are subject to.
Someone in our family bought our kid one of those talking dog toys that sings, etc. One night, my wife and I were falling asleep when we heard from the shelf, "peak-a-boo, I seeee you". Terrifying, but to this day we get a good laugh out of it.
Here's a wild theory; If you wanted to test out a worm in the Alexa network & didn't want to be detected by phoning-home, wouldn't having it do something just like this to attract media attention act as a means of proving success?
51 comments
[ 2.9 ms ] story [ 101 ms ] threadThis falls under Furby bugs level of weirdness.
http://official-furby.wikia.com/wiki/Furby_Bugs
So likely it's a false trigger on that phrase. It would be amusing if it's some kind of malware however.
My guess is that a developer replaced the "I didn't hear what you said." response with this creepy laugh on his home Alexa, then somehow accidentally pushed that build to the public.
The other explanations are a hack or disgruntled employee which seem worse.
Could be silly. Could be a veiled threat.
Other unprompted activations have been reported in abundance when Amazon installed chatbot functionality after the Alexa Prize: Alexa would start having conversations with herself in the middle of the night. One can expect bugs for newly added features, and of course the creepy ones are more likely to make the news than any random glitch.
From their perspective they don't and never will.
Come on people, this can't be real. We need to step back.
Amazon responded to the creepiness in a statement to The Verge, saying, “We’re aware of this and working to fix it.”
The Verge is not the only source that Amazon has made this statement to.
So unless this is some fake phenomenon that Amazon dreamt up to raise Alexa awareness, why do you doubt it's real?
I think anyone who thinks that this is okay, didn't have their prescribed dosage of Max Headroom re-runs.
It is just too dystopian to be real. We can't seriously be running as fast as we can off this cliff, really people?
And we got here by way of technological apathy towards the ethics issues. And .. We are still here. Everyone thinks its okay, because "entrepreneurial triumph", and so on..
We need to step back from this abyss and take a good hard look at what we're making.
Yet we all carry a far more capable mobile spying device (with audio recording, video+photo capture, and even GPS tracking all built-in) in our pockets and keep it within arms reach at all times.
At least with an Echo you can look at the traffic volume it sends through your Wifi connection to get some sense of when it's sending audio back home but you can't easily audit the traffic your phone sends, you only know what the phone and your carrier report to you.
At least with the Echo, I can be pretty sure that it's only running Amazon software, while that free game I just downloaded might be taking a photo of me right now.
Because you have reasoned that it is acceptable to not have control over who/what observes your life, in the pursuit of .. convenience and leisure ..
Yes, we have lost control over our platforms. Some of us remember a time when it really was the user in control. However, there are now generations of users who simply don't care, they just want the fancy new shit that means they don't have to get off the couch.
And, so on.
(Plus, please see the aforementioned Max Headrom, by way of levity...)
Some of us have more sense than to do that.
Also, article says it happens often as a bizarre response to requests to turn of the lights. Maybe "lights" is being interpreted as "laugh" as in a command.
But I still expect there it is some misheard utterance causing it.
Oh: it would be amusing if some malware is going around causing it.
My guess is this is in that class of defect.