Several weeks ago, terrified people topped social media with stories of a sinister, chuckling Alexa. The Amazon home-assistant was laughing without a corresponding command to chortle. These frightened folks are not publicity hounds or conspiracy theorists. Amazon acknowledged the problem and explained that Alexa might be mishearing and reacting to a presumed request to laugh. We live with speakers that can laugh on command, but somehow the creepy part is that the device laughed without being told to do so.
Objects have become our increasingly responsive companions. We can talk to and boss around everything from our fridges to our phones. Articles have chronicled toddlers ordering around adults with the same language they use to talk to their digital home companion. Techno-optimists believe that these things — A.I., chatbots, smart home devices — can solve for human inefficiencies and shortcomings, but is that possible if they can’t laugh? You may not think this is a relevant question, but I believe that in this moment of techno-laughter we are experiencing why the techno-optimists are wrong. My use of “experiencing” is intentional because I believe this realization is something that we must feel in order to understand. Alexa’s laugh paralyzes us in our profound human-ness and reminds us that objects can be effin’ scary.
What is a laugh? This question illuminates what we can gain from Alexa’s amused hiccup, but the answer is not simple. You could define a laugh by its sound or consider it the reaction to something funny. You could note that it was contagious or the best medicine. When you pay attention to these dimensions, you recognize that a laugh is both embedded and embodied. The soundwaves of laughing bounce off your eardrums, and from this perspective, you can easily call the noise emanating from Alexa a “laugh.” We hear it. But, if a laugh is a reaction to something funny, spreads amongst friends, and makes you feel better, if laughing requires and affects a body, can the snickering of disembodied code count? You could also stipulate whether laughter is voluntary or involuntary. Think about a moment you just lost it and you might understand why some people argue that all laughter is canned laughter. You simply can’t control it, and if it is automatic, then isn't it machinic? Here again, then, we find another argument for why we could accept that Alexa can laugh. If all laughter is merely canned, then Alexa’s giggle is as genuine as your Aunt Linda’s guffaw.
Let’s set aside the question of whether or not we can with any certainty call Alexa’s noise-making a “laugh” and instead focus on what we do know: An object that emits a humanesque voice is broadcasting a noise that sounds like a laugh, and the unwanted noise elicited strong reactions from people. For example, @GavinHightower tweeted: “Lying in bed about to fall asleep when Alexa on my Amazon Echo Dot lets out a very loud and creepy laugh... there’s a good chance I get murdered tonight.” You probably do not account for how Alexa’s sound has influenced your day. Suddenly, however, Hightower, and others, found Alexa loud and creepy. In a crystallizing moment, Alexa attuned listeners to a previously unvoiced tension between silence and acknowledgment. We sense that these objects are there, listening. We see them in their silence perched on desks and atop fridges, but the unrequested laugh brought before our ears the inculpatory evidence.
The laugh has reminded us that objects and environments work in mysterious and productive ways, changing us without language or logic. Even the silent and unassuming Echo Dot perched on the nightstand is influencing us. We engage in the daily, trivial use of Alexa and other digital voice services. Home assistants are unassuming in their everyday presence, but they have also made us more unassuming. Those that live with voice assistants take for granted that the “assistants” turn off the music, tell what time it is, or order toilet paper. You do not have to navigate strewn toys as you ascend a staircase, experience a tinge of burn in your thighs, and confront a pile of unwashed laundry to turn off the music upstairs. You do not have to glance at your wrist and be reminded of your dad who gave you his retirement watch before he passed away. And, thank goodness, you are no longer forced to feel embarrassed when you run into your former crush while wearing dirty sweatpants and holding a pack of Charmin in aisle four. Alexa is always present, responsive, and ready to assume the most mundane task, and in turn, every task has become mundane. Alexa is always listening. When we recognize its everyday force, or ambiance, Alexa is orchestrating banality. The unwelcomed laugh was a cry that pierced the silent routine.
Hannah Arendt wrote about banality. Arendt created the phrase “banality of evil” while covering the trial of Adolf Eichmann, an officer in the German SS who facilitated the murder of hundreds of thousands of Jewish people during the Holocaust. The path from the Holocaust to a laughing digital voice assistant may appear tenuous at best and insensitive at worst, but from trains to IBM punch cards, technology helped Eichmann carry out the Holocaust. Whereas some might argue that evil is perpetrated by a calculating mastermind, Arendt demonstrated how evil can be expertly executed by a thoughtless rule follower. It was an absence of thinking that put the banal in evil and concerned Arendt.
Keeping in mind banal thoughtlessness, I return to the question: what is a laugh? Can a laugh be banal? I say no because I believe that laughter and thinking are important counterparts. The phrase “Oh, you think that is funny?” implies this relationship. Imagine what banal laughter would sound like, and I think you might envision a noise much like Alexa’s laugh— really creepy, but perhaps more importantly, inauthentic. The laughter we covet may not be controllable or explainable, but it is conditioned by experience. We do not all laugh at the same things because we do not experience the same things. We do not think about the same things or if we do, we do not think about them in the same way. Laughing is inextricably linked to thinking even if in the moment of involuntary tittering there is a suspension of thought. In fact, it is the ability of a joke or situation to suspend thought that imbues it with comedy. Comic relief is a relief from something. If you are able to continue to think about, evaluate, and critique an ill-timed or racist joke, you will not laugh. It is the intervention into something worth interrupting that makes laughter magical. We know when someone is laughing to be polite, in other words, when they are thinking about their laughter. A Redditor’s reflection that the Alexa noise they heard “definitely sounded like a canned laugh, not like someone laughing live” demonstrates the lived phenomena of laughter, that all laughter is not canned. The joy of laughter exists in the suspension, and if there is no thought to be suspended, there is no joy, no contagion, no medicine. There is simply trivial, scary noise. Or, Arendt might say banal noise.
Journalists have mentioned Masahiro Mori’s uncanny valley in articles about Alexa’s laughter. Mori argued that humans were amused by robots until they reached a threshold of being too human without being quite human enough. He called that area of too human, not human enough the uncanny valley. In Old English cunnan meant “to know,” and thus, we could interpret un-canny as unknowing, which today we define and experience as unsettling. Current technologists want to jump the uncanny valley and build A.I. that can completely pass for human. Even in 1970 Mori wrote that some robot developers believed that robots could improve upon humans, but Mori wanted to theorize the uncanny valley in part because he wanted to understand what makes us human. Some current engineers theorize that if they improve a robot’s ability to display emotion, they can increase our affinity for humanoids. If they can make laughter more real and well-timed, they hypothesize, we will learn to love our robot leaders.
To answer Mori’s question about what makes us human, I think we can learn from our gut reaction to the Amazonian laugh. We think Alexa’s unsolicited laughter is creepy because she is not thinking. Returning to the Old English, A.I. optimists overlook that canned emotions will always be uncanny because machines will always be un-knowing. We know this in part because day-in and day-out we embody Alexa’s thoughtlessness. We feel it reducing daily moments to dullness. The unease people have experienced in coded laughter is triggering a visceral reflection on this co-created inability to think. To presume that the lack of thought is a defining characteristic of objects or to believe that the lack of thought means that something or someone is harmless is wrong. Arendt understood that a code impenetrable by thought could perpetrate horrific crimes. And, whether Alexa is ordering duct tape or laughing inexplicably, it is simply following code and asking you to bend your language and habits to its necessity. So, before condemning Amazon for buggy code or unwarranted surveillance, perhaps we should ask, who has invited this thoughtless object into our homes to make decisions and control our lights? And, we should remember, never trust a thing that can’t laugh.