It felt more like a scene from a horror movie than an example of cutting-edge technology but now, Amazon Alexa’s random laughter has been silenced.
As far as creepy tech stories go, this was a good one. Amazon Alexa’s all over the world were spontaneously bursting into laughter for seemingly no good reason and it was freaking people out.
Social media was alight with stories of un-solicited cackling laughter emanating from their voice assistance. Sometimes coming from nowhere and sometimes in response to a request for a simple task. A task that Alexa would then refuse to complete.
Amazon immediately responded, saying they were aware of the problem and putting the error down to the Echo devices: “Mistakenly hearing the phrase ‘Alexa, Laugh’.”
The issue has now been fixed.
The next Alexa update will disable the command “Alexa laugh”, which was creating false positives and replace it with the invocation “Alexa, can you laugh?” which, being a more complex sentence, is less likely to be misheard.
There are two takeaways from these changes.
One: When we are creating our own Voice First App’s it’s important to consider markedly different commands to activate different functions. The closer two commands are, the more risk of confusion and the more frustrating this will become for a user.
Two: Technology cant communicate in the same way as human’s… yet. AI and Voice Recognition are still learning (and improving) but it currently cant listen in the same way as the human ear of decipher audio information in the same way as the human brain (for example it cant read emotion or sarcasm). These devices will get better at understanding what we want but there will always be instances of mistranslation.
This creates an interesting question for programmers. Voice First technology works best when it is as close as possible to natural, human speech but at present, Alexa and her siblings struggle with the nuances of such communication. So, how do we strike the right balance between functionality, simplicity and effortless voice requests?
There is also potential that incidents such as this could negatively impact the population of Voice Assistants into the market beyond the “early adopters”. One of the most disconcerting things about this technology, for some users, is the fact they are “Always On”. This story is a reminder that Alexa is indeed always listening and constantly trying to understand the noises around her, which could put some potential users off.
The positive spin on the same angle, however, is that Alexa is doing exactly what it should be doing: Listening and learning. The more it hears, translates and understands the better the technology will become, the fewer mistakes will be made and the quicker these devices will become an essential part of everyday life.